AI/ML API provider with ten measured routes

.github/workflows/model-discovery.yml

@@ -53,6 +53,7 @@ jobs:
           # Only secrets present in repository settings become configured
           # provider credentials. Scope them to this step so checkout, setup,
           # and npm lifecycle scripts never receive provider keys.
+          AIMLAPI_API_KEY: ${{ secrets.AIMLAPI_API_KEY }}
           AINETCAFE_API_KEY: ${{ secrets.AINETCAFE_API_KEY }}
           ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
           CEREBRAS_API_KEY: ${{ secrets.CEREBRAS_API_KEY }}

README.md

@@ -255,6 +255,18 @@ Linux installations support the Codex CLI.
 
 | Picker label | Model ID | Authentication |
 | --- | --- | --- |
+| GPT-6.1 Sol (AI/ML API) | `aimlapi/gpt-6.1-sol` | AI/ML API key (`AIMLAPI_API_KEY`) |
+| GPT-6 Sol (AI/ML API) | `aimlapi/gpt-6-sol` | AI/ML API key (`AIMLAPI_API_KEY`) |
+| GPT-6 Luna (AI/ML API) | `aimlapi/gpt-6-luna` | AI/ML API key (`AIMLAPI_API_KEY`) |
+| Claude Opus 5.5 (AI/ML API) | `aimlapi/claude-opus-5.5` | AI/ML API key (`AIMLAPI_API_KEY`) |
+| Claude Sonnet 5.5 (AI/ML API) | `aimlapi/claude-sonnet-5.5` | AI/ML API key (`AIMLAPI_API_KEY`) |
+| Claude Sonnet 5 (AI/ML API) | `aimlapi/claude-sonnet-5` | AI/ML API key (`AIMLAPI_API_KEY`) |
+| Gemini 3.8 Flash (AI/ML API) | `aimlapi/gemini-3.8-flash` | AI/ML API key (`AIMLAPI_API_KEY`) |
+| DeepSeek V4.1 Flash (AI/ML API) | `aimlapi/deepseek-v4.1-flash` | AI/ML API key (`AIMLAPI_API_KEY`) |
+| GLM 5.2 (AI/ML API) | `aimlapi/glm-5.2` | AI/ML API key (`AIMLAPI_API_KEY`) |
+| Kimi K3 (AI/ML API) | `aimlapi/kimi-k3` | AI/ML API key (`AIMLAPI_API_KEY`) |
+| Grok 4.7 (AI/ML API) | `aimlapi/grok-4.7` | AI/ML API key (`AIMLAPI_API_KEY`) |
+| Qwen3.7 Max (AI/ML API) | `aimlapi/qwen3.7-max` | AI/ML API key (`AIMLAPI_API_KEY`) |
 | K2.7 Coding Highspeed (OAuth) | `kimi-oauth/kimi-for-coding-highspeed` | Existing Kimi Code CLI OAuth session |
 | K2.7 Coding (OAuth) | `kimi-oauth/kimi-for-coding` | Existing Kimi Code CLI OAuth session |
 | Kimi K3 (OAuth) | `kimi-oauth/k3` | Existing Kimi Code CLI OAuth session |

apps/macos/ModelRouterTray/Resources/PROVIDER-ICON-SOURCES.md

@@ -29,6 +29,7 @@ blue accent and DeepSeek's original blue mark are preserved.
 | NanoGPT | https://nano-gpt.com/ | https://nano-gpt.com/favicon.ico (same official diamond mark as https://nano-gpt.com/logo.png) |
 | Google Cloud Vertex AI | https://cloud.google.com/vertex-ai/ | https://www.google.com/s2/favicons?domain=cloud.google.com&sz=128 (Google mark, bundled as `google.svg`) |
 | StepFun | https://www.stepfun.com/ | https://www.stepfun.com/step_favicon.svg |
+| AI/ML API | https://aimlapi.com/ | Official hexagon mark from the AI/ML API app (`sidebar-logo.svg`) |
 
 The Z.AI, Qwen, Ollama, Cline, MiniMax, and Meta AI marks were fetched on
 2026-08-15. The Venice, Nous Research, and OpenRouter marks were fetched on

apps/macos/ModelRouterTray/Sources/IslandOverlay.swift

@@ -1085,6 +1085,7 @@ struct ProviderIcon: View {
     if providerID == "venice" { return "venice" }
     if providerID == "nousresearch" { return "nousresearch" }
     if providerID == "openrouter" { return "openrouter" }
+    if providerID == "aimlapi" { return "aimlapi" }
     if providerID == "nano-gpt" { return "nano-gpt" }
     // opencode-free plus the opencode-go API/Messages/Responses routes.
     if providerID.hasPrefix("opencode") { return "opencode-free" }
@@ -1104,7 +1105,7 @@ struct ProviderIcon: View {
   private var assetExtension: String {
     // Keyed off the asset, not the provider id, so every route sharing a mark
     // (opencode-go and friends) resolves the same file type.
-    ["github-copilot", "chutes", "google", "opencode-free", "kilo-free", "nano-gpt", "stepfun"]
+    ["aimlapi", "github-copilot", "chutes", "google", "opencode-free", "kilo-free", "nano-gpt", "stepfun"]
       .contains(assetName ?? "") ? "svg" : "png"
   }
 
@@ -1125,6 +1126,7 @@ struct ProviderIcon: View {
     if providerID == "venice" { return "Venice" }
     if providerID == "nousresearch" { return "Nous Research" }
     if providerID == "openrouter" { return "OpenRouter" }
+    if providerID == "aimlapi" { return "AI/ML API" }
     if providerID == "nano-gpt" { return "NanoGPT" }
     if providerID == "opencode-free" { return "OpenCode Free" }
     if providerID == "kilo-free" { return "Kilo Free" }

apps/macos/ModelRouterTray/Sources/ModelRouterTrayApp.swift

@@ -1624,6 +1624,7 @@ final class RouterStore: ObservableObject {
     "orca": "OrcaRouter",
     "venice": "Venice",
     "nousresearch": "Nous",
+    "aimlapi": "AI/ML API",
     "openrouter": "OpenRouter",
   ]
 

config/aimlapi/aimlapi.json

@@ -0,0 +1,24 @@
+{
+  "version": 1,
+  "providers": [
+    {
+      "id": "aimlapi",
+      "displayName": "AI/ML API",
+      "kind": "openai-compatible",
+      "ownedBy": "aimlapi",
+      "baseUrl": "https://api.aimlapi.com/v1",
+      "baseUrlEnv": "AIMLAPI_API_BASE_URL",
+      "credential": {
+        "environment": [
+          "AIMLAPI_API_KEY"
+        ],
+        "file": "aimlapi-api-key.secret",
+        "legacyFiles": [],
+        "keychainServices": [
+          "codex-router-aimlapi"
+        ],
+        "prompt": "AI/ML API key"
+      }
+    }
+  ]
+}

config/aimlapi/claude-opus-5.5.json

@@ -0,0 +1,45 @@
+{
+  "version": 1,
+  "models": [
+    {
+      "slug": "aimlapi/claude-opus-5.5",
+      "gatewayModel": "aimlapi-claude-opus-5-5",
+      "upstreamModel": "anthropic/claude-opus-5.5",
+      "provider": "aimlapi",
+      "listed": true,
+      "displayName": "Claude Opus 5.5 (AI/ML API)",
+      "description": "Anthropic Claude Opus 5.5 through AI/ML API's OpenAI-compatible surface; the full effort ladder up to max.",
+      "priority": 4,
+      "defaultEffort": "high",
+      "reasoningLevels": [
+        {
+          "effort": "low",
+          "description": "Faster reasoning"
+        },
+        {
+          "effort": "medium",
+          "description": "Balanced reasoning"
+        },
+        {
+          "effort": "high",
+          "description": "Deep reasoning"
+        },
+        {
+          "effort": "xhigh",
+          "description": "Extra deep reasoning"
+        },
+        {
+          "effort": "max",
+          "description": "Maximum reasoning depth"
+        }
+      ],
+      "contextWindow": 1000000,
+      "autoCompact": 900000,
+      "inputModalities": [
+        "text",
+        "image"
+      ],
+      "compHash": "aimlapi-claude-opus-5-5-v1"
+    }
+  ]
+}

config/aimlapi/claude-sonnet-5.5.json

@@ -0,0 +1,45 @@
+{
+  "version": 1,
+  "models": [
+    {
+      "slug": "aimlapi/claude-sonnet-5.5",
+      "gatewayModel": "aimlapi-claude-sonnet-5-5",
+      "upstreamModel": "anthropic/claude-sonnet-5.5",
+      "provider": "aimlapi",
+      "listed": true,
+      "displayName": "Claude Sonnet 5.5 (AI/ML API)",
+      "description": "Anthropic Claude Sonnet 5.5 through AI/ML API; the current Sonnet, full effort ladder up to max.",
+      "priority": 5,
+      "defaultEffort": "high",
+      "reasoningLevels": [
+        {
+          "effort": "low",
+          "description": "Faster reasoning"
+        },
+        {
+          "effort": "medium",
+          "description": "Balanced reasoning"
+        },
+        {
+          "effort": "high",
+          "description": "Deep reasoning"
+        },
+        {
+          "effort": "xhigh",
+          "description": "Extra deep reasoning"
+        },
+        {
+          "effort": "max",
+          "description": "Maximum reasoning depth"
+        }
+      ],
+      "contextWindow": 1000000,
+      "autoCompact": 900000,
+      "inputModalities": [
+        "text",
+        "image"
+      ],
+      "compHash": "aimlapi-claude-sonnet-5-5-v1"
+    }
+  ]
+}

config/aimlapi/claude-sonnet-5.json

@@ -0,0 +1,45 @@
+{
+  "version": 1,
+  "models": [
+    {
+      "slug": "aimlapi/claude-sonnet-5",
+      "gatewayModel": "aimlapi-claude-sonnet-5",
+      "upstreamModel": "anthropic/claude-sonnet-5",
+      "provider": "aimlapi",
+      "listed": true,
+      "displayName": "Claude Sonnet 5 (AI/ML API)",
+      "description": "Anthropic Claude Sonnet 5 through AI/ML API; balanced coding model with the full effort ladder.",
+      "priority": 6,
+      "defaultEffort": "high",
+      "reasoningLevels": [
+        {
+          "effort": "low",
+          "description": "Faster reasoning"
+        },
+        {
+          "effort": "medium",
+          "description": "Balanced reasoning"
+        },
+        {
+          "effort": "high",
+          "description": "Deep reasoning"
+        },
+        {
+          "effort": "xhigh",
+          "description": "Extra deep reasoning"
+        },
+        {
+          "effort": "max",
+          "description": "Maximum reasoning depth"
+        }
+      ],
+      "contextWindow": 1000000,
+      "autoCompact": 900000,
+      "inputModalities": [
+        "text",
+        "image"
+      ],
+      "compHash": "aimlapi-claude-sonnet-5-v1"
+    }
+  ]
+}

config/aimlapi/deepseek-v4.1-flash.json

@@ -0,0 +1,37 @@
+{
+  "version": 1,
+  "models": [
+    {
+      "slug": "aimlapi/deepseek-v4.1-flash",
+      "gatewayModel": "aimlapi-deepseek-v4-1-flash",
+      "upstreamModel": "deepseek/deepseek-v4.1-flash",
+      "provider": "aimlapi",
+      "listed": true,
+      "displayName": "DeepSeek V4.1 Flash (AI/ML API)",
+      "description": "DeepSeek V4.1 Flash through AI/ML API; fast reasoning with tool calls and image understanding.",
+      "priority": 8,
+      "defaultEffort": "high",
+      "reasoningLevels": [
+        {
+          "effort": "low",
+          "description": "Faster reasoning"
+        },
+        {
+          "effort": "medium",
+          "description": "Balanced reasoning"
+        },
+        {
+          "effort": "high",
+          "description": "Deep reasoning"
+        }
+      ],
+      "contextWindow": 1048576,
+      "autoCompact": 900000,
+      "inputModalities": [
+        "text",
+        "image"
+      ],
+      "compHash": "aimlapi-deepseek-v4-1-flash-v1"
+    }
+  ]
+}

config/aimlapi/gemini-3.8-flash.json

@@ -0,0 +1,41 @@
+{
+  "version": 1,
+  "models": [
+    {
+      "slug": "aimlapi/gemini-3.8-flash",
+      "gatewayModel": "aimlapi-gemini-3-8-flash",
+      "upstreamModel": "google/gemini-3.8-flash",
+      "provider": "aimlapi",
+      "listed": true,
+      "displayName": "Gemini 3.8 Flash (AI/ML API)",
+      "description": "Google Gemini 3.8 Flash through AI/ML API; 1M context, text and image input. Rejects xhigh, accepts max.",
+      "priority": 7,
+      "defaultEffort": "high",
+      "reasoningLevels": [
+        {
+          "effort": "low",
+          "description": "Faster reasoning"
+        },
+        {
+          "effort": "medium",
+          "description": "Balanced reasoning"
+        },
+        {
+          "effort": "high",
+          "description": "Deep reasoning"
+        },
+        {
+          "effort": "max",
+          "description": "Maximum reasoning depth"
+        }
+      ],
+      "contextWindow": 1048576,
+      "autoCompact": 900000,
+      "inputModalities": [
+        "text",
+        "image"
+      ],
+      "compHash": "aimlapi-gemini-3-8-flash-v1"
+    }
+  ]
+}

config/aimlapi/glm-5.2.json

@@ -0,0 +1,48 @@
+{
+  "version": 1,
+  "models": [
+    {
+      "slug": "aimlapi/glm-5.2",
+      "gatewayModel": "aimlapi-glm-5-2",
+      "upstreamModel": "zhipu/glm-5.2",
+      "provider": "aimlapi",
+      "listed": true,
+      "displayName": "GLM 5.2 (AI/ML API)",
+      "description": "Zhipu GLM 5.2 through AI/ML API; text only on this route, and the only model here that accepts every effort level.",
+      "priority": 9,
+      "defaultEffort": "high",
+      "reasoningLevels": [
+        {
+          "effort": "minimal",
+          "description": "Minimal reasoning"
+        },
+        {
+          "effort": "low",
+          "description": "Faster reasoning"
+        },
+        {
+          "effort": "medium",
+          "description": "Balanced reasoning"
+        },
+        {
+          "effort": "high",
+          "description": "Deep reasoning"
+        },
+        {
+          "effort": "xhigh",
+          "description": "Extra deep reasoning"
+        },
+        {
+          "effort": "max",
+          "description": "Maximum reasoning depth"
+        }
+      ],
+      "contextWindow": 1000000,
+      "autoCompact": 900000,
+      "inputModalities": [
+        "text"
+      ],
+      "compHash": "aimlapi-glm-5-2-v1"
+    }
+  ]
+}