{ "version": "2026.08.25", "generatedAt": "2026-08-25T22:26:05.935Z", "tier": "monthly", "counts": { "platforms": 26, "models": 269, "enabledModels": 269, "embeddings": 16, "enabledEmbeddings": 16, "transcriptionModels": 4, "videoModels": 0, "quirks": 30 }, "platforms": [ { "id": "agnes", "name": "Agnes AI" }, { "id": "aihorde", "name": "AI Horde" }, { "id": "ainative", "name": "AINative Studio" }, { "id": "aion", "name": "Aion Labs" }, { "id": "bazaarlink", "name": "BazaarLink" }, { "id": "cerebras", "name": "Cerebras" }, { "id": "cloudflare", "name": "Cloudflare Workers AI" }, { "id": "cohere", "name": "Cohere" }, { "id": "google", "name": "Google AI Studio" }, { "id": "groq", "name": "Groq" }, { "id": "huggingface", "name": "HuggingFace Router" }, { "id": "kilo", "name": "Kilo Gateway" }, { "id": "llm7", "name": "LLM7" }, { "id": "mistral", "name": "Mistral" }, { "id": "nara", "name": "NaraRouter" }, { "id": "navy", "name": "NavyAI" }, { "id": "nvidia", "name": "NVIDIA NIM" }, { "id": "ollama", "name": "Ollama Cloud" }, { "id": "opencode", "name": "OpenCode Zen" }, { "id": "openrouter", "name": "OpenRouter" }, { "id": "ovh", "name": "OVH AI Endpoints" }, { "id": "reka", "name": "Reka" }, { "id": "requesty", "name": "Requesty" }, { "id": "routeway", "name": "Routeway" }, { "id": "sealion", "name": "SEA-LION" }, { "id": "zhipu", "name": "Zhipu AI" } ], "models": [ { "platform": "google", "modelId": "gemini-3.5-flash", "displayName": "Gemini 3.5 Flash", "intelligenceRank": 3, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 10, "rpd": 20, "tpm": 250000, "tpd": null }, "monthlyTokenBudget": "~3M", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [ { "slug": "gemini-thinking-token-room", "title": "Thinking tokens use output cap", "body": "Gemini Flash can spend hidden thinking tokens inside maxOutputTokens. Avoid tiny max_tokens values or set thinkingBudget to 0 when provider support is available.", "severity": "info" } ] }, { "platform": "huggingface", "modelId": "Qwen/Qwen3-Coder-Next", "displayName": "Qwen3-Coder Next (HF)", "intelligenceRank": 3, "speedRank": 9, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~1-3M", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "hf-tiny-credit", "title": "Small $0.10/month routed credit", "body": "HuggingFace Inference Providers grants only ~$0.10/month of routed credit on the free tier (PRO is $2/month). Enough for light experimentation; exhausts quickly. Credits apply only to HF-routed requests.", "severity": "warning" } ] }, { "platform": "huggingface", "modelId": "moonshotai/Kimi-K2.6", "displayName": "Kimi K2.6 (HF)", "intelligenceRank": 3, "speedRank": 9, "sizeLabel": "Frontier", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~1-3M", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "hf-tiny-credit", "title": "Small $0.10/month routed credit", "body": "HuggingFace Inference Providers grants only ~$0.10/month of routed credit on the free tier (PRO is $2/month). Enough for light experimentation; exhausts quickly. Credits apply only to HF-routed requests.", "severity": "warning" } ] }, { "platform": "huggingface", "modelId": "deepseek-ai/DeepSeek-V4-Flash", "displayName": "DeepSeek V4 Flash (HF)", "intelligenceRank": 4, "speedRank": 9, "sizeLabel": "Frontier", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~1-3M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "hf-tiny-credit", "title": "Small $0.10/month routed credit", "body": "HuggingFace Inference Providers grants only ~$0.10/month of routed credit on the free tier (PRO is $2/month). Enough for light experimentation; exhausts quickly. Credits apply only to HF-routed requests.", "severity": "warning" } ] }, { "platform": "opencode", "modelId": "deepseek-v4-flash-free", "displayName": "DeepSeek V4 Flash Free (OpenCode Zen)", "intelligenceRank": 4, "speedRank": 4, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "promo (trial)", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "zen-promo-roster", "title": "Limited-time promo, roster rotates", "body": "OpenCode Zen free models are explicitly limited-time promotional access (\"available for a limited time\" per the docs), not a recurring quota. The roster rotates: qwen3.6-plus and minimax-m3 promos already ended. Expect any row here to die without notice; prompts/outputs may be used for model improvement.", "severity": "warning" } ] }, { "platform": "ovh", "modelId": "Qwen3.5-397B-A17B", "displayName": "Qwen3.5 397B (OVH)", "intelligenceRank": 5, "speedRank": 9, "sizeLabel": "Frontier", "limits": { "rpm": 2, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 2/min per IP", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "ovh-anon-trickle", "title": "Anonymous tier is 2 req/min", "body": "OVH AI Endpoints anonymous mode is documented at 2 req/min per IP per model (observed even stricter across models). The 400 req/min authenticated tier requires a Public Cloud project with a payment method, so the catalog ships the keyless path. Treat as a breadth/fallback tier, not a throughput tier.", "severity": "warning" }, { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" } ] }, { "platform": "cerebras", "modelId": "gpt-oss-120b", "displayName": "GPT-OSS 120B (Cerebras)", "intelligenceRank": 6, "speedRank": 1, "sizeLabel": "Large", "limits": { "rpm": 5, "rpd": null, "tpm": 30000, "tpd": 1000000 }, "monthlyTokenBudget": "~30M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "cloudflare", "modelId": "@cf/openai/gpt-oss-120b", "displayName": "GPT-OSS 120B (CF)", "intelligenceRank": 6, "speedRank": 11, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~18-45M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" }, { "slug": "reasoning-token-room", "title": "Needs token room", "body": "This route can spend hidden reasoning tokens before visible output. Avoid tiny max_tokens values or the request can finish by length with empty content.", "severity": "info" } ] }, { "platform": "groq", "modelId": "groq/compound", "displayName": "Compound (Groq)", "intelligenceRank": 6, "speedRank": 2, "sizeLabel": "Large", "limits": { "rpm": 30, "rpd": 250, "tpm": 70000, "tpd": null }, "monthlyTokenBudget": "~6M", "contextWindow": 131072, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [] }, { "platform": "groq", "modelId": "openai/gpt-oss-120b", "displayName": "GPT-OSS 120B (Groq)", "intelligenceRank": 6, "speedRank": 2, "sizeLabel": "Large", "limits": { "rpm": 30, "rpd": 1000, "tpm": 8000, "tpd": 200000 }, "monthlyTokenBudget": "~6M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "ollama", "modelId": "gpt-oss:120b", "displayName": "GPT-OSS 120B (Ollama)", "intelligenceRank": 6, "speedRank": 9, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~10-20M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "ovh", "modelId": "gpt-oss-120b", "displayName": "GPT-OSS 120B (OVH)", "intelligenceRank": 6, "speedRank": 9, "sizeLabel": "Large", "limits": { "rpm": 2, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 2/min per IP", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "ovh-anon-trickle", "title": "Anonymous tier is 2 req/min", "body": "OVH AI Endpoints anonymous mode is documented at 2 req/min per IP per model (observed even stricter across models). The 400 req/min authenticated tier requires a Public Cloud project with a payment method, so the catalog ships the keyless path. Treat as a breadth/fallback tier, not a throughput tier.", "severity": "warning" }, { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" } ] }, { "platform": "cerebras", "modelId": "zai-glm-4.7", "displayName": "GLM-4.7 (Cerebras)", "intelligenceRank": 7, "speedRank": 1, "sizeLabel": "Large", "limits": { "rpm": 5, "rpd": null, "tpm": 30000, "tpd": 1000000 }, "monthlyTokenBudget": "~30M", "contextWindow": 8192, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "cloudflare", "modelId": "@cf/qwen/qwen3-30b-a3b-fp8", "displayName": "Qwen3 30B-A3B fp8 (CF)", "intelligenceRank": 7, "speedRank": 11, "sizeLabel": "Medium", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~18-45M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ] }, { "platform": "opencode", "modelId": "nemotron-3-ultra-free", "displayName": "Nemotron 3 Ultra Free (OpenCode Zen)", "intelligenceRank": 7, "speedRank": 4, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "promo (trial)", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "zen-promo-roster", "title": "Limited-time promo, roster rotates", "body": "OpenCode Zen free models are explicitly limited-time promotional access (\"available for a limited time\" per the docs), not a recurring quota. The roster rotates: qwen3.6-plus and minimax-m3 promos already ended. Expect any row here to die without notice; prompts/outputs may be used for model improvement.", "severity": "warning" }, { "slug": "zen-serves-ultra-fast", "title": "Zen serves the 550B fast", "body": "OpenCode Zen serves nemotron-3-ultra in ~2s with working tool calls where the OpenRouter route hangs — the live-verified path for this model.", "severity": "info" } ] }, { "platform": "openrouter", "modelId": "nvidia/nemotron-3-ultra-550b-a55b:free", "displayName": "Nemotron 3 Ultra 550B (free, slow)", "intelligenceRank": 7, "speedRank": 11, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~6M", "contextWindow": 1000000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "or-free-cap-account-wide", "title": "Daily :free cap is account-wide", "body": "OpenRouter’s :free daily cap (50/day, or 1000/day once you have ever bought $10 of credits) is shared across ALL :free models on the account, not per model. Per-row rpd values here are therefore optimistic; the router’s cooldown handling absorbs the shared 429s.", "severity": "info" } ] }, { "platform": "cloudflare", "modelId": "@cf/deepseek-ai/deepseek-r1-distill-qwen-32b", "displayName": "DeepSeek R1 Distill Qwen 32B (CF)", "intelligenceRank": 9, "speedRank": 11, "sizeLabel": "Medium", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~3-5M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ] }, { "platform": "cloudflare", "modelId": "@cf/nvidia/nemotron-3-120b-a12b", "displayName": "Nemotron 3 120B (CF)", "intelligenceRank": 9, "speedRank": 11, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~5-10M", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ] }, { "platform": "cloudflare", "modelId": "@cf/zai-org/glm-4.7-flash", "displayName": "GLM-4.7 Flash (CF)", "intelligenceRank": 10, "speedRank": 11, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~18-45M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ] }, { "platform": "opencode", "modelId": "big-pickle", "displayName": "Big Pickle (OpenCode Zen, stealth)", "intelligenceRank": 10, "speedRank": 4, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "promo (trial)", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "zen-promo-roster", "title": "Limited-time promo, roster rotates", "body": "OpenCode Zen free models are explicitly limited-time promotional access (\"available for a limited time\" per the docs), not a recurring quota. The roster rotates: qwen3.6-plus and minimax-m3 promos already ended. Expect any row here to die without notice; prompts/outputs may be used for model improvement.", "severity": "warning" } ] }, { "platform": "google", "modelId": "gemini-3-flash-preview", "displayName": "Gemini 3 Flash Preview", "intelligenceRank": 11, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 10, "rpd": 20, "tpm": 250000, "tpd": null }, "monthlyTokenBudget": "~3M", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [ { "slug": "gemini-thinking-token-room", "title": "Thinking tokens use output cap", "body": "Gemini Flash can spend hidden thinking tokens inside maxOutputTokens. Avoid tiny max_tokens values or set thinkingBudget to 0 when provider support is available.", "severity": "info" } ] }, { "platform": "cloudflare", "modelId": "@cf/meta/llama-4-scout-17b-16e-instruct", "displayName": "Llama 4 Scout (CF)", "intelligenceRank": 12, "speedRank": 11, "sizeLabel": "Medium", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~18-45M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ] }, { "platform": "cohere", "modelId": "command-a-reasoning-08-2025", "displayName": "Command A Reasoning (08-2025)", "intelligenceRank": 13, "speedRank": 11, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": 33, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~1-2M", "contextWindow": 256000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "cohere-reasoning-output", "title": "May emit reasoning prose", "body": "Cohere Command A Reasoning can spend output budget explaining its thought process on terse prompts. Use a larger max_tokens cap or a stricter no-reasoning instruction when exact short answers are required.", "severity": "warning" } ] }, { "platform": "opencode", "modelId": "north-mini-code-free", "displayName": "North Mini Code Free (OpenCode Zen)", "intelligenceRank": 13, "speedRank": 4, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "promo (trial)", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "zen-promo-roster", "title": "Limited-time promo, roster rotates", "body": "OpenCode Zen free models are explicitly limited-time promotional access (\"available for a limited time\" per the docs), not a recurring quota. The roster rotates: qwen3.6-plus and minimax-m3 promos already ended. Expect any row here to die without notice; prompts/outputs may be used for model improvement.", "severity": "warning" } ] }, { "platform": "kilo", "modelId": "stepfun/step-3.7-flash:free", "displayName": "StepFun Step 3.7 Flash (Kilo)", "intelligenceRank": 14, "speedRank": 3, "sizeLabel": "Medium", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 200/hr per IP", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" } ] }, { "platform": "mistral", "modelId": "mistral-large-latest", "displayName": "Mistral Large 3", "intelligenceRank": 14, "speedRank": 8, "sizeLabel": "Medium", "limits": { "rpm": 2, "rpd": null, "tpm": 500000, "tpd": null }, "monthlyTokenBudget": "~50-100M", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "mistral", "modelId": "mistral-medium-latest", "displayName": "Mistral Medium 3.5", "intelligenceRank": 14, "speedRank": 8, "sizeLabel": "Large", "limits": { "rpm": 2, "rpd": null, "tpm": 500000, "tpd": null }, "monthlyTokenBudget": "~50-100M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "mistral", "modelId": "mistral-small-latest", "displayName": "Mistral Small 4", "intelligenceRank": 14, "speedRank": 8, "sizeLabel": "Medium", "limits": { "rpm": 2, "rpd": null, "tpm": 500000, "tpd": null }, "monthlyTokenBudget": "~50-100M", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "opencode", "modelId": "mimo-v2.5-free", "displayName": "MiMo-V2.5 Free (OpenCode Zen)", "intelligenceRank": 14, "speedRank": 4, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "promo (trial)", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "zen-promo-roster", "title": "Limited-time promo, roster rotates", "body": "OpenCode Zen free models are explicitly limited-time promotional access (\"available for a limited time\" per the docs), not a recurring quota. The roster rotates: qwen3.6-plus and minimax-m3 promos already ended. Expect any row here to die without notice; prompts/outputs may be used for model improvement.", "severity": "warning" } ] }, { "platform": "llm7", "modelId": "codestral-latest", "displayName": "Codestral (LLM7)", "intelligenceRank": 16, "speedRank": 8, "sizeLabel": "Small", "limits": { "rpm": 100, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~2M (60-100/hr)", "contextWindow": 32000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" } ] }, { "platform": "mistral", "modelId": "codestral-latest", "displayName": "Codestral", "intelligenceRank": 16, "speedRank": 6, "sizeLabel": "Small", "limits": { "rpm": 2, "rpd": null, "tpm": 500000, "tpd": null }, "monthlyTokenBudget": "~50-100M", "contextWindow": 256000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "mistral", "modelId": "devstral-latest", "displayName": "Devstral", "intelligenceRank": 16, "speedRank": 8, "sizeLabel": "Medium", "limits": { "rpm": 2, "rpd": null, "tpm": 500000, "tpd": null }, "monthlyTokenBudget": "~50-100M", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "cloudflare", "modelId": "@cf/meta/llama-3.3-70b-instruct-fp8-fast", "displayName": "Llama 3.3 70B fp8-fast (CF)", "intelligenceRank": 17, "speedRank": 11, "sizeLabel": "Medium", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~18-45M", "contextWindow": 24000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ] }, { "platform": "nvidia", "modelId": "meta/llama-3.1-70b-instruct", "displayName": "Llama 3.1 70B (NV)", "intelligenceRank": 17, "speedRank": 6, "sizeLabel": "Medium", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ] }, { "platform": "nvidia", "modelId": "meta/llama-3.3-70b-instruct", "displayName": "Llama 3.3 70B (NV)", "intelligenceRank": 17, "speedRank": 6, "sizeLabel": "Medium", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ] }, { "platform": "ovh", "modelId": "Meta-Llama-3_3-70B-Instruct", "displayName": "Llama 3.3 70B (OVH)", "intelligenceRank": 17, "speedRank": 9, "sizeLabel": "Medium", "limits": { "rpm": 2, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 2/min per IP", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "ovh-anon-trickle", "title": "Anonymous tier is 2 req/min", "body": "OVH AI Endpoints anonymous mode is documented at 2 req/min per IP per model (observed even stricter across models). The 400 req/min authenticated tier requires a Public Cloud project with a payment method, so the catalog ships the keyless path. Treat as a breadth/fallback tier, not a throughput tier.", "severity": "warning" }, { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" } ] }, { "platform": "google", "modelId": "gemini-3.1-flash-lite", "displayName": "Gemini 3.1 Flash-Lite", "intelligenceRank": 18, "speedRank": 3, "sizeLabel": "Large", "limits": { "rpm": 15, "rpd": 20, "tpm": 250000, "tpd": null }, "monthlyTokenBudget": "~3M", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "groq", "modelId": "groq/compound-mini", "displayName": "Compound Mini (Groq)", "intelligenceRank": 18, "speedRank": 2, "sizeLabel": "Medium", "limits": { "rpm": 30, "rpd": 250, "tpm": 70000, "tpd": null }, "monthlyTokenBudget": "~6M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "groq", "modelId": "openai/gpt-oss-20b", "displayName": "GPT-OSS 20B (Groq)", "intelligenceRank": 18, "speedRank": 2, "sizeLabel": "Medium", "limits": { "rpm": 30, "rpd": 1000, "tpm": 8000, "tpd": 200000 }, "monthlyTokenBudget": "~6M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "groq", "modelId": "openai/gpt-oss-safeguard-20b", "displayName": "GPT-OSS Safeguard 20B (Groq)", "intelligenceRank": 18, "speedRank": 2, "sizeLabel": "Medium", "limits": { "rpm": 30, "rpd": 1000, "tpm": 8000, "tpd": 200000 }, "monthlyTokenBudget": "~6M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "cloudflare", "modelId": "@cf/openai/gpt-oss-20b", "displayName": "GPT-OSS 20B (CF)", "intelligenceRank": 18, "speedRank": 11, "sizeLabel": "Medium", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~18-45M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" }, { "slug": "reasoning-token-room", "title": "Needs token room", "body": "This route can spend hidden reasoning tokens before visible output. Avoid tiny max_tokens values or the request can finish by length with empty content.", "severity": "info" } ] }, { "platform": "zhipu", "modelId": "glm-4.7-flash", "displayName": "GLM-4.7 Flash", "intelligenceRank": 18, "speedRank": 4, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": 1000000 }, "monthlyTokenBudget": "~30M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "google", "modelId": "gemma-4-31b-it", "displayName": "Gemma 4 31B IT", "intelligenceRank": 19, "speedRank": 4, "sizeLabel": "Large", "limits": { "rpm": 15, "rpd": 1000, "tpm": 250000, "tpd": null }, "monthlyTokenBudget": "~30M", "contextWindow": 32768, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [] }, { "platform": "nvidia", "modelId": "google/gemma-4-31b-it", "displayName": "Gemma 4 31B (NV)", "intelligenceRank": 19, "speedRank": 9, "sizeLabel": "Large", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ] }, { "platform": "cerebras", "modelId": "gemma-4-31b", "displayName": "Gemma 4 31B (Cerebras)", "intelligenceRank": 19, "speedRank": 1, "sizeLabel": "Large", "limits": { "rpm": 5, "rpd": null, "tpm": 30000, "tpd": 1000000 }, "monthlyTokenBudget": "~30M", "contextWindow": 32768, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [] }, { "platform": "openrouter", "modelId": "google/gemma-4-31b-it:free", "displayName": "Gemma 4 31B (free)", "intelligenceRank": 19, "speedRank": 9, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~6M", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "or-free-cap-account-wide", "title": "Daily :free cap is account-wide", "body": "OpenRouter’s :free daily cap (50/day, or 1000/day once you have ever bought $10 of credits) is shared across ALL :free models on the account, not per model. Per-row rpd values here are therefore optimistic; the router’s cooldown handling absorbs the shared 429s.", "severity": "info" } ] }, { "platform": "ovh", "modelId": "Qwen3-Coder-30B-A3B-Instruct", "displayName": "Qwen3-Coder 30B (OVH)", "intelligenceRank": 19, "speedRank": 9, "sizeLabel": "Medium", "limits": { "rpm": 2, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 2/min per IP", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "ovh-anon-trickle", "title": "Anonymous tier is 2 req/min", "body": "OVH AI Endpoints anonymous mode is documented at 2 req/min per IP per model (observed even stricter across models). The 400 req/min authenticated tier requires a Public Cloud project with a payment method, so the catalog ships the keyless path. Treat as a breadth/fallback tier, not a throughput tier.", "severity": "warning" }, { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" } ] }, { "platform": "google", "modelId": "gemini-2.5-flash", "displayName": "Gemini 2.5 Flash", "intelligenceRank": 20, "speedRank": 5, "sizeLabel": "Medium", "limits": { "rpm": 10, "rpd": 20, "tpm": 250000, "tpd": null }, "monthlyTokenBudget": "~3M", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "google", "modelId": "gemma-4-26b-a4b-it", "displayName": "Gemma 4 26B IT", "intelligenceRank": 20, "speedRank": 4, "sizeLabel": "Large", "limits": { "rpm": 15, "rpd": 1000, "tpm": 250000, "tpd": null }, "monthlyTokenBudget": "~30M", "contextWindow": 32768, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [] }, { "platform": "mistral", "modelId": "magistral-medium-latest", "displayName": "Magistral Medium", "intelligenceRank": 21, "speedRank": 8, "sizeLabel": "Large", "limits": { "rpm": 2, "rpd": null, "tpm": 500000, "tpd": null }, "monthlyTokenBudget": "~50-100M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "zhipu", "modelId": "glm-4.6v-flash", "displayName": "GLM-4.6V Flash", "intelligenceRank": 21, "speedRank": 4, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~30M", "contextWindow": 131072, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [ { "slug": "zhipu-shared-key", "title": "Works with existing Zhipu key", "body": "glm-4.6v-flash is listed Free on Z.AI and answers 200 with the existing bigmodel.cn key; vision and structured tool calls both live-verified.", "severity": "info" } ] }, { "platform": "cloudflare", "modelId": "@cf/google/gemma-4-26b-a4b-it", "displayName": "Gemma 4 26B-A4B it (CF)", "intelligenceRank": 22, "speedRank": 11, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~10-20M", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ] }, { "platform": "kilo", "modelId": "nvidia/nemotron-3-super-120b-a12b:free", "displayName": "Nemotron 3 Super 120B (Kilo)", "intelligenceRank": 22, "speedRank": 9, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~2-3M (200/hr)", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" } ] }, { "platform": "nvidia", "modelId": "nvidia/nemotron-3-nano-30b-a3b", "displayName": "Nemotron 3 Nano 30B (NV)", "intelligenceRank": 22, "speedRank": 9, "sizeLabel": "Medium", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ] }, { "platform": "nvidia", "modelId": "nvidia/nemotron-3-super-120b-a12b", "displayName": "Nemotron 3 Super 120B (NV)", "intelligenceRank": 22, "speedRank": 9, "sizeLabel": "Large", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ] }, { "platform": "ollama", "modelId": "gemma4:31b", "displayName": "Gemma 4 31B (Ollama)", "intelligenceRank": 22, "speedRank": 10, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~20-30M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "openrouter", "modelId": "google/gemma-4-26b-a4b-it:free", "displayName": "Gemma 4 26B-A4B (free)", "intelligenceRank": 22, "speedRank": 9, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~6M", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "or-free-cap-account-wide", "title": "Daily :free cap is account-wide", "body": "OpenRouter’s :free daily cap (50/day, or 1000/day once you have ever bought $10 of credits) is shared across ALL :free models on the account, not per model. Per-row rpd values here are therefore optimistic; the router’s cooldown handling absorbs the shared 429s.", "severity": "info" } ] }, { "platform": "openrouter", "modelId": "nvidia/nemotron-3-super-120b-a12b:free", "displayName": "Nemotron 3 Super 120B (free)", "intelligenceRank": 22, "speedRank": 9, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~6M", "contextWindow": 1000000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "or-free-cap-account-wide", "title": "Daily :free cap is account-wide", "body": "OpenRouter’s :free daily cap (50/day, or 1000/day once you have ever bought $10 of credits) is shared across ALL :free models on the account, not per model. Per-row rpd values here are therefore optimistic; the router’s cooldown handling absorbs the shared 429s.", "severity": "info" } ] }, { "platform": "openrouter", "modelId": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free", "displayName": "Nemotron 3 Nano 30B Reasoning (free)", "intelligenceRank": 23, "speedRank": 9, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~6M", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "or-free-cap-account-wide", "title": "Daily :free cap is account-wide", "body": "OpenRouter’s :free daily cap (50/day, or 1000/day once you have ever bought $10 of credits) is shared across ALL :free models on the account, not per model. Per-row rpd values here are therefore optimistic; the router’s cooldown handling absorbs the shared 429s.", "severity": "info" } ] }, { "platform": "openrouter", "modelId": "poolside/laguna-xs-2.1:free", "displayName": "Poolside Laguna XS 2.1 (free)", "intelligenceRank": 26, "speedRank": 10, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~6M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "or-free-cap-account-wide", "title": "Daily :free cap is account-wide", "body": "OpenRouter’s :free daily cap (50/day, or 1000/day once you have ever bought $10 of credits) is shared across ALL :free models on the account, not per model. Per-row rpd values here are therefore optimistic; the router’s cooldown handling absorbs the shared 429s.", "severity": "info" }, { "slug": "reasoning-token-room", "title": "Needs token room", "body": "This route can spend 90+ hidden reasoning tokens before visible output. Avoid tiny max_tokens values or the request can finish by length with empty content.", "severity": "info" } ] }, { "platform": "cohere", "modelId": "command-r-08-2024", "displayName": "Command R (08-2024)", "intelligenceRank": 25, "speedRank": 11, "sizeLabel": "Small", "limits": { "rpm": 20, "rpd": 33, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~1-2M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "google", "modelId": "gemini-2.5-flash-lite", "displayName": "Gemini 2.5 Flash-Lite", "intelligenceRank": 26, "speedRank": 3, "sizeLabel": "Medium", "limits": { "rpm": 15, "rpd": 20, "tpm": 250000, "tpd": null }, "monthlyTokenBudget": "~3M", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "cohere", "modelId": "command-a-03-2025", "displayName": "Command-A (03-2025)", "intelligenceRank": 27, "speedRank": 11, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": 33, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~1-2M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "cohere", "modelId": "command-r-plus-08-2024", "displayName": "Command R+ (08-2024)", "intelligenceRank": 27, "speedRank": 11, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": 33, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~1-2M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "mistral", "modelId": "ministral-8b-latest", "displayName": "Ministral 3 8B", "intelligenceRank": 28, "speedRank": 8, "sizeLabel": "Small", "limits": { "rpm": 2, "rpd": null, "tpm": 500000, "tpd": null }, "monthlyTokenBudget": "~50-100M", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "cloudflare", "modelId": "@cf/ibm-granite/granite-4.0-h-micro", "displayName": "Granite 4.0 H Micro (CF)", "intelligenceRank": 29, "speedRank": 11, "sizeLabel": "Small", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~5-10M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ] }, { "platform": "cohere", "modelId": "command-a-plus-05-2026", "displayName": "Command A+ (05-2026)", "intelligenceRank": 12, "speedRank": 11, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": 33, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~1-2M", "contextWindow": 131072, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "agnes", "modelId": "agnes-2.0-flash", "displayName": "Agnes 2.0 Flash", "intelligenceRank": 17, "speedRank": 6, "sizeLabel": "Medium", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · promo $0/token", "contextWindow": 262144, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "reka", "modelId": "reka-flash", "displayName": "Reka Flash", "intelligenceRank": 12, "speedRank": 6, "sizeLabel": "Medium", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~50M", "contextWindow": 65536, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "reka", "modelId": "reka-edge-2603", "displayName": "Reka Edge", "intelligenceRank": 20, "speedRank": 4, "sizeLabel": "Small", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~100M", "contextWindow": 16384, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "cloudflare", "modelId": "@cf/black-forest-labs/flux-1-schnell", "displayName": "FLUX.1 [schnell]", "modality": "image", "intelligenceRank": 203, "speedRank": 203, "sizeLabel": "12B", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "", "contextWindow": null, "enabled": true, "supportsVision": false, "supportsTools": false, "mediaNote": "Shared 10k neurons/day", "quirks": [] }, { "platform": "cloudflare", "modelId": "@cf/stabilityai/stable-diffusion-xl-base-1.0", "displayName": "Stable Diffusion XL", "modality": "image", "intelligenceRank": 204, "speedRank": 204, "sizeLabel": "SDXL", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "", "contextWindow": null, "enabled": true, "supportsVision": false, "supportsTools": false, "mediaNote": "Shared 10k neurons/day", "quirks": [] }, { "platform": "cloudflare", "modelId": "@cf/myshell-ai/melotts", "displayName": "MeloTTS", "modality": "audio", "intelligenceRank": 211, "speedRank": 211, "sizeLabel": "Small", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "", "contextWindow": null, "enabled": true, "supportsVision": false, "supportsTools": false, "mediaNote": "MP3 output - multilingual", "quirks": [] }, { "platform": "google", "modelId": "gemini-2.5-flash-preview-tts", "displayName": "Gemini 2.5 Flash TTS", "modality": "audio", "intelligenceRank": 214, "speedRank": 214, "sizeLabel": "Flash", "limits": { "rpm": 3, "rpd": 15, "tpm": null, "tpd": null }, "monthlyTokenBudget": "", "contextWindow": null, "enabled": true, "supportsVision": false, "supportsTools": false, "mediaNote": "WAV/PCM - 30+ voices", "quirks": [] }, { "platform": "zhipu", "modelId": "glm-4.5-flash", "displayName": "GLM-4.5 Flash", "intelligenceRank": 22, "speedRank": 4, "sizeLabel": "Medium", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": 1000000 }, "monthlyTokenBudget": "~30M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "cohere", "modelId": "command-a-vision-07-2025", "displayName": "Command A Vision (07-2025)", "intelligenceRank": 15, "speedRank": 10, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": 33, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~1-2M", "contextWindow": 128000, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [] }, { "platform": "openrouter", "modelId": "cohere/north-mini-code:free", "displayName": "North Mini Code (free)", "intelligenceRank": 21, "speedRank": 8, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": 50, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~6M", "contextWindow": 256000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "or-free-cap-account-wide", "title": "Daily :free cap is account-wide", "body": "OpenRouter’s :free daily cap (50/day, or 1000/day once you have ever bought $10 of credits) is shared across ALL :free models on the account, not per model. Per-row rpd values here are therefore optimistic; the router’s cooldown handling absorbs the shared 429s.", "severity": "info" } ] }, { "platform": "openrouter", "modelId": "nvidia/nemotron-3.5-content-safety:free", "displayName": "Nemotron 3.5 Content Safety (free)", "intelligenceRank": 31, "speedRank": 9, "sizeLabel": "Small", "limits": { "rpm": 20, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~4M", "contextWindow": 128000, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "or-free-cap-account-wide", "title": "Daily :free cap is account-wide", "body": "OpenRouter’s :free daily cap (50/day, or 1000/day once you have ever bought $10 of credits) is shared across ALL :free models on the account, not per model. Per-row rpd values here are therefore optimistic; the router’s cooldown handling absorbs the shared 429s.", "severity": "info" } ] }, { "platform": "groq", "modelId": "qwen/qwen3.6-27b", "displayName": "Qwen3.6 27B (Groq)", "intelligenceRank": 17, "speedRank": 2, "sizeLabel": "Medium", "limits": { "rpm": 60, "rpd": 1000, "tpm": 6000, "tpd": 500000 }, "monthlyTokenBudget": "~15M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "ovh", "modelId": "gpt-oss-20b", "displayName": "GPT-OSS 20B (OVH)", "intelligenceRank": 14, "speedRank": 10, "sizeLabel": "Medium", "limits": { "rpm": 2, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 2/min per IP", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "ovh-anon-trickle", "title": "Anonymous tier is 2 req/min", "body": "OVH AI Endpoints anonymous mode is documented at 2 req/min per IP per model (observed even stricter across models). The 400 req/min authenticated tier requires a Public Cloud project with a payment method, so the catalog ships the keyless path. Treat as a breadth/fallback tier, not a throughput tier.", "severity": "warning" }, { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" }, { "slug": "reasoning-token-room", "title": "Needs token room", "body": "Some free routes spend hidden reasoning tokens before visible output. Avoid tiny max_tokens values or requests can finish by length with empty content.", "severity": "info" } ] }, { "platform": "routeway", "modelId": "llama-3.3-70b-instruct:free", "displayName": "Llama 3.3 70B Instruct (Routeway free)", "intelligenceRank": 10, "speedRank": 7, "sizeLabel": "Large", "limits": { "rpm": 5, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · ~5 rpm · 200 rpd", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "routeway", "modelId": "llama-3.2-3b-instruct:free", "displayName": "Llama 3.2 3B Instruct (Routeway free)", "intelligenceRank": 18, "speedRank": 10, "sizeLabel": "Small", "limits": { "rpm": 5, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · ~5 rpm · 200 rpd", "contextWindow": 16000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [], "premiumSince": "2026-07-19T18:44:22.000Z" }, { "platform": "bazaarlink", "modelId": "auto:free", "displayName": "BazaarLink Auto (free router)", "intelligenceRank": 8, "speedRank": 7, "sizeLabel": "Medium", "limits": { "rpm": 10, "rpd": 150, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · auto route", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "ainative", "modelId": "qwen3-32b", "displayName": "Qwen3 32B (AINative)", "intelligenceRank": 7, "speedRank": 7, "sizeLabel": "Medium", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 10M tok/mo (claimed)", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "ainative", "modelId": "llama-4-maverick", "displayName": "Llama 4 Maverick (AINative)", "intelligenceRank": 8, "speedRank": 7, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 10M tok/mo (claimed)", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "ainative", "modelId": "qwen3-14b", "displayName": "Qwen3 14B (AINative)", "intelligenceRank": 11, "speedRank": 8, "sizeLabel": "Medium", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 10M tok/mo (claimed)", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "ainative", "modelId": "qwen3-8b", "displayName": "Qwen3 8B (AINative)", "intelligenceRank": 14, "speedRank": 9, "sizeLabel": "Small", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 10M tok/mo (claimed)", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "aihorde", "modelId": "aphrodite/TheDrummer/Cydonia-24B-v4.3", "displayName": "Cydonia 24B v4.3 (AI Horde)", "intelligenceRank": 178, "speedRank": 205, "sizeLabel": "24B", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "Kudos queue (no token cap)", "contextWindow": 8192, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "aihorde-anon-slow", "title": "Free volunteer queue, slow", "body": "AI Horde routes to volunteer-run workers through a priority queue, so latency is seconds to minutes, not the sub-second of hosted providers. The anonymous key 0000000000 runs at the lowest priority; register a free key at aihorde.net for higher priority. The provider uses a 120s timeout.", "severity": "warning" }, { "slug": "aihorde-no-tools", "title": "No tool calling", "body": "AI Horde's OpenAI-compatible proxy does not support function/tool calling. The provider drops tools, tool_choice and parallel_tool_calls so a tool-using request still completes as plain chat instead of failing.", "severity": "info" }, { "slug": "aihorde-usage-estimated", "title": "Usage is kudos; tokens estimated", "body": "The proxy returns usage as {\"kudos\": N} with no token counts, and rejects max_tokens below 16 and a non-array stop. The AIHordeProvider normalizes the request (floors max_tokens, wraps stop) and synthesizes prompt/completion token estimates so analytics and savings math aren't zero.", "severity": "info" }, { "slug": "aihorde-roster-rotates", "title": "Roster + context depend on online workers", "body": "Model availability changes as volunteer workers come and go, so a listed model can be temporarily unserved. The effective context window is set by the worker (often 4-8K), not the model's native maximum.", "severity": "info" }, { "slug": "aihorde-quality-uneven", "title": "Uneven quality", "body": "Output quality varies by worker and quantization, and some workers append template or instruction text after the answer. Best treated as free fallback capacity, not a primary model.", "severity": "warning" } ] }, { "platform": "nvidia", "modelId": "nvidia/llama-3.3-nemotron-super-49b-v1.5", "displayName": "Nemotron Super 49B v1.5 (NV)", "intelligenceRank": 9, "speedRank": 8, "sizeLabel": "Large", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ] }, { "platform": "nvidia", "modelId": "stepfun-ai/step-3.7-flash", "displayName": "Step 3.7 Flash (NV)", "intelligenceRank": 13, "speedRank": 7, "sizeLabel": "Medium", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 65536, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ] }, { "platform": "nvidia", "modelId": "minimaxai/minimax-m3", "displayName": "MiniMax M3 (NV)", "intelligenceRank": 2, "speedRank": 9, "sizeLabel": "Frontier", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 196608, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ] }, { "platform": "cohere", "modelId": "c4ai-aya-expanse-32b", "displayName": "Aya Expanse 32B", "intelligenceRank": 22, "speedRank": 10, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": 33, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~1-2M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "cohere", "modelId": "c4ai-aya-vision-32b", "displayName": "Aya Vision 32B", "intelligenceRank": 20, "speedRank": 10, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": 33, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~1-2M", "contextWindow": 16384, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [] }, { "platform": "cohere", "modelId": "north-mini-code-1-0", "displayName": "North Mini Code", "intelligenceRank": 21, "speedRank": 11, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": 33, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~1-2M", "contextWindow": 256000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "cohere", "modelId": "command-r7b-12-2024", "displayName": "Command R7B (12-2024)", "intelligenceRank": 30, "speedRank": 6, "sizeLabel": "Small", "limits": { "rpm": 20, "rpd": 33, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~1-2M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "groq", "modelId": "allam-2-7b", "displayName": "ALLaM 2 7B (Groq)", "intelligenceRank": 32, "speedRank": 2, "sizeLabel": "Small", "limits": { "rpm": 30, "rpd": 1000, "tpm": 6000, "tpd": 500000 }, "monthlyTokenBudget": "~15M", "contextWindow": 4096, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "kilo", "modelId": "nvidia/nemotron-3-ultra-550b-a55b:free", "displayName": "Nemotron 3 Ultra 550B (Kilo)", "intelligenceRank": 7, "speedRank": 9, "sizeLabel": "Frontier", "contextWindow": 1000000, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" } ], "monthlyTokenBudget": "free · 200/hr per IP (trial)", "enabled": true, "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null } }, { "platform": "kilo", "modelId": "poolside/laguna-xs-2.1:free", "displayName": "Poolside Laguna XS 2.1 (Kilo)", "intelligenceRank": 16, "speedRank": 4, "sizeLabel": "Medium", "contextWindow": 262144, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" }, { "slug": "reasoning-token-room", "title": "Needs token room", "body": "This route can spend 90+ hidden reasoning tokens before visible output. Avoid tiny max_tokens values or the request can finish by length with empty content.", "severity": "info" } ], "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 200/hr per IP", "enabled": true }, { "platform": "kilo", "modelId": "kilo-auto/free", "displayName": "Kilo Auto Free", "intelligenceRank": 18, "speedRank": 10, "sizeLabel": "Large", "contextWindow": 256000, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" }, { "slug": "kilo-auto-token-room", "title": "Auto route needs token room", "body": "Kilo auto/free can route to reasoning models that spend 100+ internal reasoning tokens. Avoid tiny max_tokens or a request can finish by length with empty content.", "severity": "info" } ], "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 200/hr per IP", "enabled": true }, { "platform": "kilo", "modelId": "openrouter/free", "displayName": "Free Router (Kilo)", "intelligenceRank": 18, "speedRank": 10, "sizeLabel": "Large", "contextWindow": 200000, "supportsVision": true, "supportsTools": true, "quirks": [ { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" } ], "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 200/hr per IP", "enabled": true }, { "platform": "kilo", "modelId": "cohere/north-mini-code:free", "displayName": "North Mini Code (Kilo)", "intelligenceRank": 21, "speedRank": 8, "sizeLabel": "Medium", "contextWindow": 256000, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" } ], "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 200/hr per IP", "enabled": true }, { "platform": "kilo", "modelId": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free", "displayName": "Nemotron 3 Nano Omni Reasoning (Kilo)", "intelligenceRank": 23, "speedRank": 9, "sizeLabel": "Medium", "contextWindow": 256000, "supportsVision": true, "supportsTools": false, "quirks": [ { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" } ], "monthlyTokenBudget": "free · 200/hr per IP (trial)", "enabled": true, "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null } }, { "platform": "kilo", "modelId": "nvidia/nemotron-3.5-content-safety:free", "displayName": "Nemotron 3.5 Content Safety (Kilo)", "intelligenceRank": 31, "speedRank": 9, "sizeLabel": "Small", "contextWindow": 128000, "supportsVision": true, "supportsTools": false, "quirks": [ { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" }, { "slug": "safety-classifier-output", "title": "Safety classifier output", "body": "This route returns moderation/safety labels rather than normal assistant prose. Use it as a guardrail model, not a primary chat model.", "severity": "info" } ], "monthlyTokenBudget": "free · 200/hr per IP (trial)", "enabled": true, "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null } }, { "platform": "nara", "modelId": "mistral-large", "displayName": "Mistral Large 3 (NaraRouter)", "intelligenceRank": 14, "speedRank": 5, "sizeLabel": "Medium", "limits": { "rpm": 10, "rpd": null, "tpm": null, "tpd": 7000000 }, "monthlyTokenBudget": "free · 7M/day shared", "contextWindow": 252000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "nara-free-plan", "title": "Daily free quota", "body": "NaraRouter free access requires a no-card account key plus Telegram channel/link verification. The free plan is documented at 7M tokens/day and 10 req/min, with daily reset.", "severity": "info" } ] }, { "platform": "nara", "modelId": "mistral-medium-3-5", "displayName": "Mistral Medium 3.5 (NaraRouter)", "intelligenceRank": 14, "speedRank": 5, "sizeLabel": "Large", "limits": { "rpm": 10, "rpd": null, "tpm": null, "tpd": 7000000 }, "monthlyTokenBudget": "free · 7M/day shared", "contextWindow": 256000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [ { "slug": "nara-free-plan", "title": "Daily free quota", "body": "NaraRouter free access requires a no-card account key plus Telegram channel/link verification. The free plan is documented at 7M tokens/day and 10 req/min, with daily reset.", "severity": "info" } ] }, { "platform": "mistral", "modelId": "mistral-code-latest", "displayName": "Mistral Code", "intelligenceRank": 13, "speedRank": 8, "sizeLabel": "Medium", "limits": { "rpm": 2, "rpd": null, "tpm": 500000, "tpd": null }, "monthlyTokenBudget": "~50-100M", "contextWindow": 256000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "mistral", "modelId": "mistral-code-agent-latest", "displayName": "Mistral Code Agent", "intelligenceRank": 13, "speedRank": 8, "sizeLabel": "Medium", "limits": { "rpm": 2, "rpd": null, "tpm": 500000, "tpd": null }, "monthlyTokenBudget": "~50-100M", "contextWindow": 256000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "mistral", "modelId": "devstral-medium-latest", "displayName": "Devstral Medium", "intelligenceRank": 14, "speedRank": 8, "sizeLabel": "Large", "limits": { "rpm": 2, "rpd": null, "tpm": 500000, "tpd": null }, "monthlyTokenBudget": "~50-100M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "mistral", "modelId": "magistral-small-latest", "displayName": "Magistral Small", "intelligenceRank": 22, "speedRank": 8, "sizeLabel": "Small", "limits": { "rpm": 2, "rpd": null, "tpm": 500000, "tpd": null }, "monthlyTokenBudget": "~50-100M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "mistral", "modelId": "ministral-14b-latest", "displayName": "Ministral 14B", "intelligenceRank": 27, "speedRank": 8, "sizeLabel": "Small", "limits": { "rpm": 2, "rpd": null, "tpm": 500000, "tpd": null }, "monthlyTokenBudget": "~50-100M", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "mistral", "modelId": "mistral-vibe-cli-fast", "displayName": "Mistral Vibe CLI Fast", "intelligenceRank": 16, "speedRank": 7, "sizeLabel": "Medium", "limits": { "rpm": 2, "rpd": null, "tpm": 500000, "tpd": null }, "monthlyTokenBudget": "~50-100M", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "ovh", "modelId": "Qwen3-32B", "displayName": "Qwen3 32B (OVH)", "intelligenceRank": 19, "speedRank": 9, "sizeLabel": "Medium", "limits": { "rpm": 2, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 2/min per IP", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "ovh-anon-trickle", "title": "Anonymous tier is 2 req/min", "body": "OVH AI Endpoints anonymous mode is documented at 2 req/min per IP per model (observed even stricter across models). The 400 req/min authenticated tier requires a Public Cloud project with a payment method, so the catalog ships the keyless path. Treat as a breadth/fallback tier, not a throughput tier.", "severity": "warning" }, { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" }, { "slug": "ovh-qwen-think-trace", "title": "May emit thinking trace", "body": "OVH Qwen reasoning routes can include blocks or reasoning prose. Strip or suppress the thinking trace when a plain answer is required.", "severity": "warning" } ] }, { "platform": "ovh", "modelId": "Qwen3.6-27B", "displayName": "Qwen3.6 27B (OVH)", "intelligenceRank": 17, "speedRank": 9, "sizeLabel": "Medium", "limits": { "rpm": 2, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 2/min per IP", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "ovh-anon-trickle", "title": "Anonymous tier is 2 req/min", "body": "OVH AI Endpoints anonymous mode is documented at 2 req/min per IP per model (observed even stricter across models). The 400 req/min authenticated tier requires a Public Cloud project with a payment method, so the catalog ships the keyless path. Treat as a breadth/fallback tier, not a throughput tier.", "severity": "warning" }, { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" }, { "slug": "ovh-qwen-think-trace", "title": "May emit thinking trace", "body": "OVH Qwen reasoning routes can include blocks or reasoning prose. Strip or suppress the thinking trace when a plain answer is required.", "severity": "warning" }, { "slug": "reasoning-token-room", "title": "Needs token room", "body": "Some free routes spend hidden reasoning tokens before visible output. Avoid tiny max_tokens values or requests can finish by length with empty content.", "severity": "info" } ] }, { "platform": "ovh", "modelId": "Qwen2.5-VL-72B-Instruct", "displayName": "Qwen2.5 VL 72B (OVH)", "intelligenceRank": 21, "speedRank": 9, "sizeLabel": "Medium", "limits": { "rpm": 2, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 2/min per IP", "contextWindow": 32768, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [ { "slug": "ovh-anon-trickle", "title": "Anonymous tier is 2 req/min", "body": "OVH AI Endpoints anonymous mode is documented at 2 req/min per IP per model (observed even stricter across models). The 400 req/min authenticated tier requires a Public Cloud project with a payment method, so the catalog ships the keyless path. Treat as a breadth/fallback tier, not a throughput tier.", "severity": "warning" }, { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" } ] }, { "platform": "ovh", "modelId": "Mistral-Small-3.2-24B-Instruct-2506", "displayName": "Mistral Small 3.2 24B (OVH)", "intelligenceRank": 23, "speedRank": 9, "sizeLabel": "Medium", "limits": { "rpm": 2, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 2/min per IP", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "ovh-anon-trickle", "title": "Anonymous tier is 2 req/min", "body": "OVH AI Endpoints anonymous mode is documented at 2 req/min per IP per model (observed even stricter across models). The 400 req/min authenticated tier requires a Public Cloud project with a payment method, so the catalog ships the keyless path. Treat as a breadth/fallback tier, not a throughput tier.", "severity": "warning" }, { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" } ] }, { "platform": "ovh", "modelId": "Mistral-Nemo-Instruct-2407", "displayName": "Mistral Nemo (OVH)", "intelligenceRank": 28, "speedRank": 9, "sizeLabel": "Small", "limits": { "rpm": 2, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 2/min per IP", "contextWindow": 128000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "ovh-anon-trickle", "title": "Anonymous tier is 2 req/min", "body": "OVH AI Endpoints anonymous mode is documented at 2 req/min per IP per model (observed even stricter across models). The 400 req/min authenticated tier requires a Public Cloud project with a payment method, so the catalog ships the keyless path. Treat as a breadth/fallback tier, not a throughput tier.", "severity": "warning" }, { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" } ] }, { "platform": "ovh", "modelId": "Mistral-7B-Instruct-v0.3", "displayName": "Mistral 7B Instruct v0.3 (OVH)", "intelligenceRank": 31, "speedRank": 9, "sizeLabel": "Small", "limits": { "rpm": 2, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 2/min per IP", "contextWindow": 65536, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "ovh-anon-trickle", "title": "Anonymous tier is 2 req/min", "body": "OVH AI Endpoints anonymous mode is documented at 2 req/min per IP per model (observed even stricter across models). The 400 req/min authenticated tier requires a Public Cloud project with a payment method, so the catalog ships the keyless path. Treat as a breadth/fallback tier, not a throughput tier.", "severity": "warning" }, { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" } ] }, { "platform": "ovh", "modelId": "Qwen3Guard-Gen-8B", "displayName": "Qwen3Guard Gen 8B (OVH safety)", "intelligenceRank": 31, "speedRank": 9, "sizeLabel": "Small", "limits": { "rpm": 2, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 2/min per IP", "contextWindow": 32768, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "ovh-anon-trickle", "title": "Anonymous tier is 2 req/min", "body": "OVH AI Endpoints anonymous mode is documented at 2 req/min per IP per model (observed even stricter across models). The 400 req/min authenticated tier requires a Public Cloud project with a payment method, so the catalog ships the keyless path. Treat as a breadth/fallback tier, not a throughput tier.", "severity": "warning" }, { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" }, { "slug": "safety-classifier-output", "title": "Safety classifier output", "body": "This route returns moderation/safety labels rather than normal assistant prose. Use it as a guardrail model, not a primary chat model.", "severity": "info" } ] }, { "platform": "ovh", "modelId": "Qwen3Guard-Gen-0.6B", "displayName": "Qwen3Guard Gen 0.6B (OVH safety)", "intelligenceRank": 31, "speedRank": 9, "sizeLabel": "Small", "limits": { "rpm": 2, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 2/min per IP", "contextWindow": 32768, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "ovh-anon-trickle", "title": "Anonymous tier is 2 req/min", "body": "OVH AI Endpoints anonymous mode is documented at 2 req/min per IP per model (observed even stricter across models). The 400 req/min authenticated tier requires a Public Cloud project with a payment method, so the catalog ships the keyless path. Treat as a breadth/fallback tier, not a throughput tier.", "severity": "warning" }, { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" }, { "slug": "safety-classifier-output", "title": "Safety classifier output", "body": "This route returns moderation/safety labels rather than normal assistant prose. Use it as a guardrail model, not a primary chat model.", "severity": "info" } ] }, { "platform": "nvidia", "modelId": "nvidia/nemotron-nano-12b-v2-vl", "displayName": "Nemotron Nano 12B VL (NV)", "intelligenceRank": 28, "speedRank": 9, "sizeLabel": "Medium", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 128000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ] }, { "platform": "aion", "modelId": "aion-labs/aion-3.0", "displayName": "Aion 3.0", "intelligenceRank": 10, "speedRank": 6, "sizeLabel": "Large", "limits": { "rpm": 15, "rpd": null, "tpm": 20000, "tpd": 20000 }, "monthlyTokenBudget": "free · 20k tok/day", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "aion", "modelId": "aion-labs/aion-3.0-mini", "displayName": "Aion 3.0 Mini", "intelligenceRank": 13, "speedRank": 5, "sizeLabel": "Medium", "limits": { "rpm": 15, "rpd": null, "tpm": 20000, "tpd": 20000 }, "monthlyTokenBudget": "free · 20k tok/day", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "aion", "modelId": "aion-labs/aion-2.5", "displayName": "Aion 2.5", "intelligenceRank": 13, "speedRank": 6, "sizeLabel": "Large", "limits": { "rpm": 15, "rpd": null, "tpm": 20000, "tpd": 20000 }, "monthlyTokenBudget": "free · 20k tok/day", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "aion", "modelId": "aion-labs/aion-2.0", "displayName": "Aion 2.0", "intelligenceRank": 14, "speedRank": 7, "sizeLabel": "Large", "limits": { "rpm": 15, "rpd": null, "tpm": 20000, "tpd": 20000 }, "monthlyTokenBudget": "free · 20k tok/day", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "aion", "modelId": "aion-labs/aion-rp-llama-3.1-8b", "displayName": "Aion-RP Llama 3.1 8B", "intelligenceRank": 18, "speedRank": 4, "sizeLabel": "Small", "limits": { "rpm": 15, "rpd": null, "tpm": 20000, "tpd": 20000 }, "monthlyTokenBudget": "free · 20k tok/day", "contextWindow": 32768, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "requesty", "modelId": "nvidia/nemotron-3-ultra-550b-a55b", "displayName": "Nemotron 3 Ultra 550B (Requesty)", "intelligenceRank": 7, "speedRank": 9, "sizeLabel": "Frontier", "limits": { "rpm": null, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 200 rpd", "contextWindow": 1048576, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "requesty", "modelId": "mistral/leanstral-1-5", "displayName": "Leanstral 1.5 (Requesty)", "intelligenceRank": 10, "speedRank": 7, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 200 rpd", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "requesty-non-greedy-sampling", "title": "Requires non-greedy sampling", "body": "Requesty rejected this route when temperature was 0 with greedy sampling. Use a nonzero temperature or provider-compatible sampling parameters.", "severity": "info" } ] }, { "platform": "requesty", "modelId": "poolside/laguna-m.1", "displayName": "Poolside Laguna M.1 (Requesty)", "intelligenceRank": 13, "speedRank": 8, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 200 rpd", "contextWindow": 32768, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "reasoning-token-room", "title": "Needs token room", "body": "Some free routes spend hidden reasoning tokens before visible output. Avoid tiny max_tokens values or requests can finish by length with empty content.", "severity": "info" } ] }, { "platform": "requesty", "modelId": "google/gemma-4-31b-it", "displayName": "Gemma 4 31B (Requesty)", "intelligenceRank": 19, "speedRank": 9, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 200 rpd", "contextWindow": 262144, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "requesty", "modelId": "nvidia/nemotron-3-super-120b-a12b", "displayName": "Nemotron 3 Super 120B (Requesty)", "intelligenceRank": 22, "speedRank": 9, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 200 rpd", "contextWindow": 1048576, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "requesty", "modelId": "nvidia/nemotron-3-nano-30b-a3b", "displayName": "Nemotron 3 Nano 30B (Requesty)", "intelligenceRank": 23, "speedRank": 9, "sizeLabel": "Medium", "limits": { "rpm": null, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 200 rpd", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "requesty", "modelId": "nvidia/nemotron-3.5-content-safety", "displayName": "Nemotron 3.5 Content Safety (Requesty)", "intelligenceRank": 31, "speedRank": 9, "sizeLabel": "Small", "limits": { "rpm": null, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 200 rpd", "contextWindow": 131072, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [ { "slug": "safety-classifier-output", "title": "Safety classifier output", "body": "This route returns moderation/safety labels rather than normal assistant prose. Use it as a guardrail model, not a primary chat model.", "severity": "info" } ] }, { "platform": "routeway", "modelId": "step-3.7-flash:free", "displayName": "StepFun Step 3.7 Flash (Routeway free)", "intelligenceRank": 10, "speedRank": 7, "sizeLabel": "Frontier", "limits": { "rpm": 5, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · ~5 rpm · 200 rpd", "contextWindow": 256000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-5.4", "displayName": "GPT 5.4 (NavyAI)", "intelligenceRank": 4, "speedRank": 10, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 33333 }, "monthlyTokenBudget": "~1M/month shared · 4.5x", "contextWindow": 1050000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-5.4-mini", "displayName": "GPT 5.4 Mini (NavyAI)", "intelligenceRank": 4, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 93750 }, "monthlyTokenBudget": "~2.8M/month shared · 1.6x", "contextWindow": 400000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-5.4-nano", "displayName": "GPT 5.4 Nano (NavyAI)", "intelligenceRank": 4, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 115384 }, "monthlyTokenBudget": "~3.5M/month shared · 1.3x", "contextWindow": 400000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-5.3-codex", "displayName": "GPT 5.3 Codex (NavyAI)", "intelligenceRank": 4, "speedRank": 8, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 42857 }, "monthlyTokenBudget": "~1.3M/month shared · 3.5x", "contextWindow": 400000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-5.2", "displayName": "GPT 5.2 (NavyAI)", "intelligenceRank": 4, "speedRank": 8, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 42857 }, "monthlyTokenBudget": "~1.3M/month shared · 3.5x", "contextWindow": 400000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-5.1", "displayName": "GPT 5.1 (NavyAI)", "intelligenceRank": 4, "speedRank": 8, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 60000 }, "monthlyTokenBudget": "~1.8M/month shared · 2.5x", "contextWindow": 400000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-5", "displayName": "GPT 5 (NavyAI)", "intelligenceRank": 4, "speedRank": 8, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 60000 }, "monthlyTokenBudget": "~1.8M/month shared · 2.5x", "contextWindow": 400000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-5-mini", "displayName": "GPT 5 Mini (NavyAI)", "intelligenceRank": 4, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 83333 }, "monthlyTokenBudget": "~2.5M/month shared · 1.8x", "contextWindow": 400000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-5-nano", "displayName": "GPT 5 Nano (NavyAI)", "intelligenceRank": 4, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 115384 }, "monthlyTokenBudget": "~3.5M/month shared · 1.3x", "contextWindow": 400000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-5-search-api", "displayName": "GPT 5 Search API (NavyAI)", "intelligenceRank": 4, "speedRank": 8, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 60000 }, "monthlyTokenBudget": "~1.8M/month shared · 2.5x", "contextWindow": 400000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-4o", "displayName": "GPT 4o (NavyAI)", "intelligenceRank": 13, "speedRank": 7, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 60000 }, "monthlyTokenBudget": "~1.8M/month shared · 2.5x", "contextWindow": 128000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-4o-mini", "displayName": "GPT 4o Mini (NavyAI)", "intelligenceRank": 13, "speedRank": 5, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 125000 }, "monthlyTokenBudget": "~3.8M/month shared · 1.2x", "contextWindow": 128000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-4.1", "displayName": "GPT 4.1 (NavyAI)", "intelligenceRank": 13, "speedRank": 7, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 83333 }, "monthlyTokenBudget": "~2.5M/month shared · 1.8x", "contextWindow": 1047576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-4.1-mini", "displayName": "GPT 4.1 Mini (NavyAI)", "intelligenceRank": 13, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 115384 }, "monthlyTokenBudget": "~3.5M/month shared · 1.3x", "contextWindow": 1047576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-4.1-nano", "displayName": "GPT 4.1 Nano (NavyAI)", "intelligenceRank": 13, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 125000 }, "monthlyTokenBudget": "~3.8M/month shared · 1.2x", "contextWindow": 1047576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-3.5-turbo", "displayName": "GPT 3.5 Turbo (NavyAI)", "intelligenceRank": 24, "speedRank": 6, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 107142 }, "monthlyTokenBudget": "~3.2M/month shared · 1.4x", "contextWindow": 16385, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "o4-mini", "displayName": "O4 Mini (NavyAI)", "intelligenceRank": 4, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 93750 }, "monthlyTokenBudget": "~2.8M/month shared · 1.6x", "contextWindow": 200000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "o3-mini", "displayName": "O3 Mini (NavyAI)", "intelligenceRank": 4, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 93750 }, "monthlyTokenBudget": "~2.8M/month shared · 1.6x", "contextWindow": 200000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "o3", "displayName": "O3 (NavyAI)", "intelligenceRank": 4, "speedRank": 8, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 60000 }, "monthlyTokenBudget": "~1.8M/month shared · 2.5x", "contextWindow": 200000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gemini-3.1-flash-lite", "displayName": "Gemini 3.1 Flash Lite (NavyAI)", "intelligenceRank": 18, "speedRank": 5, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 75000 }, "monthlyTokenBudget": "~2.3M/month shared · 2x", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gemini-3.1-flash-lite-thinking", "displayName": "Gemini 3.1 Flash Lite Thinking (NavyAI)", "intelligenceRank": 6, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 75000 }, "monthlyTokenBudget": "~2.3M/month shared · 2x", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gemini-3.1-flash-image", "displayName": "Gemini 3.1 Flash Image (NavyAI)", "intelligenceRank": 6, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 3947 }, "monthlyTokenBudget": "~118K/month shared · 38x", "contextWindow": 131072, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [], "premiumSince": "2026-07-23T22:58:14.000Z" }, { "platform": "navy", "modelId": "gemini-3-flash-preview", "displayName": "Gemini 3 Flash Preview (NavyAI)", "intelligenceRank": 11, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 75000 }, "monthlyTokenBudget": "~2.3M/month shared · 2x", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gemini-3-flash-preview-thinking", "displayName": "Gemini 3 Flash Preview Thinking (NavyAI)", "intelligenceRank": 6, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 75000 }, "monthlyTokenBudget": "~2.3M/month shared · 2x", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gemini-2.5-flash", "displayName": "Gemini 2.5 Flash (NavyAI)", "intelligenceRank": 20, "speedRank": 5, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 100000 }, "monthlyTokenBudget": "~3M/month shared · 1.5x", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gemini-2.5-flash-thinking", "displayName": "Gemini 2.5 Flash Thinking (NavyAI)", "intelligenceRank": 17, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 100000 }, "monthlyTokenBudget": "~3M/month shared · 1.5x", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gemini-2.5-flash-lite", "displayName": "Gemini 2.5 Flash Lite (NavyAI)", "intelligenceRank": 26, "speedRank": 5, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 125000 }, "monthlyTokenBudget": "~3.8M/month shared · 1.2x", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gemini-2.5-flash-image", "displayName": "Gemini 2.5 Flash Image (NavyAI)", "intelligenceRank": 17, "speedRank": 5, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 3947 }, "monthlyTokenBudget": "~118K/month shared · 38x", "contextWindow": 32768, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [] }, { "platform": "navy", "modelId": "gemma-4-31b-it", "displayName": "Gemma 4 31B IT (NavyAI)", "intelligenceRank": 19, "speedRank": 5, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 262144, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gemma-4-26b-a4b-it", "displayName": "Gemma 4 26B A4b IT (NavyAI)", "intelligenceRank": 20, "speedRank": 5, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 262144, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "llama-3.1-8b-instruct", "displayName": "Llama 3.1 8B Instruct (NavyAI)", "intelligenceRank": 22, "speedRank": 5, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "llama-3.3-70b-instruct", "displayName": "Llama 3.3 70B Instruct (NavyAI)", "intelligenceRank": 17, "speedRank": 6, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "deepseek-v4-flash", "displayName": "Deepseek V4 Flash (NavyAI)", "intelligenceRank": 6, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 1048576, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "deepseek-v4-pro", "displayName": "Deepseek V4 Pro (NavyAI)", "intelligenceRank": 4, "speedRank": 10, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 71428 }, "monthlyTokenBudget": "~2.1M/month shared · 2.1x", "contextWindow": 1048576, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "deepseek-chat", "displayName": "Deepseek Chat (NavyAI)", "intelligenceRank": 24, "speedRank": 6, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "deepseek-reasoner", "displayName": "Deepseek Reasoner (NavyAI)", "intelligenceRank": 24, "speedRank": 6, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 125000 }, "monthlyTokenBudget": "~3.8M/month shared · 1.2x", "contextWindow": null, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "navy", "modelId": "deepseek-v3.2", "displayName": "Deepseek V3.2 (NavyAI)", "intelligenceRank": 13, "speedRank": 7, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 163840, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "grok-4.3", "displayName": "Grok 4.3 (NavyAI)", "intelligenceRank": 4, "speedRank": 8, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 50000 }, "monthlyTokenBudget": "~1.5M/month shared · 3x", "contextWindow": 1000000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "grok-4.20-reasoning", "displayName": "Grok 4.20 Reasoning (NavyAI)", "intelligenceRank": 4, "speedRank": 10, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 33333 }, "monthlyTokenBudget": "~1M/month shared · 4.5x", "contextWindow": 2000000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "grok-4.20-non-reasoning", "displayName": "Grok 4.20 Non Reasoning (NavyAI)", "intelligenceRank": 4, "speedRank": 10, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 33333 }, "monthlyTokenBudget": "~1M/month shared · 4.5x", "contextWindow": 2000000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "grok-4.1-fast-reasoning", "displayName": "Grok 4.1 Fast Reasoning (NavyAI)", "intelligenceRank": 6, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 100000 }, "monthlyTokenBudget": "~3M/month shared · 1.5x", "contextWindow": 2000000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "grok-4.1-fast-non-reasoning", "displayName": "Grok 4.1 Fast Non Reasoning (NavyAI)", "intelligenceRank": 6, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 100000 }, "monthlyTokenBudget": "~3M/month shared · 1.5x", "contextWindow": 2000000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "grok-code-fast-1", "displayName": "Grok Code Fast 1 (NavyAI)", "intelligenceRank": 18, "speedRank": 5, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 100000 }, "monthlyTokenBudget": "~3M/month shared · 1.5x", "contextWindow": 256000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "grok-4", "displayName": "Grok 4 (NavyAI)", "intelligenceRank": 6, "speedRank": 10, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 15000 }, "monthlyTokenBudget": "~450K/month shared · 10x", "contextWindow": 256000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "grok-4-fast-reasoning", "displayName": "Grok 4 Fast Reasoning (NavyAI)", "intelligenceRank": 6, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 100000 }, "monthlyTokenBudget": "~3M/month shared · 1.5x", "contextWindow": 2000000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "grok-4-fast-non-reasoning", "displayName": "Grok 4 Fast Non Reasoning (NavyAI)", "intelligenceRank": 6, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 100000 }, "monthlyTokenBudget": "~3M/month shared · 1.5x", "contextWindow": 2000000, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "codestral-2508", "displayName": "Codestral 2508 (NavyAI)", "intelligenceRank": 17, "speedRank": 6, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 125000 }, "monthlyTokenBudget": "~3.8M/month shared · 1.2x", "contextWindow": 256000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "codestral-latest", "displayName": "Codestral Latest (NavyAI)", "intelligenceRank": 16, "speedRank": 8, "sizeLabel": "Small", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 125000 }, "monthlyTokenBudget": "~3.8M/month shared · 1.2x", "contextWindow": 256000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "mistral-small-2603", "displayName": "Mistral Small 2603 (NavyAI)", "intelligenceRank": 17, "speedRank": 6, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 262144, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "mistral-small-latest", "displayName": "Mistral Small Latest (NavyAI)", "intelligenceRank": 14, "speedRank": 8, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 262144, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "mistral-medium-3-5", "displayName": "Mistral Medium 3 5 (NavyAI)", "intelligenceRank": 14, "speedRank": 5, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 18750 }, "monthlyTokenBudget": "~563K/month shared · 8x", "contextWindow": 262144, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "mistral-medium-2508", "displayName": "Mistral Medium 2508 (NavyAI)", "intelligenceRank": 13, "speedRank": 7, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 60000 }, "monthlyTokenBudget": "~1.8M/month shared · 2.5x", "contextWindow": 128000, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [] }, { "platform": "navy", "modelId": "mistral-medium-latest", "displayName": "Mistral Medium Latest (NavyAI)", "intelligenceRank": 14, "speedRank": 8, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 18750 }, "monthlyTokenBudget": "~563K/month shared · 8x", "contextWindow": 128000, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [] }, { "platform": "navy", "modelId": "mistral-large-2512", "displayName": "Mistral Large 2512 (NavyAI)", "intelligenceRank": 6, "speedRank": 8, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 75000 }, "monthlyTokenBudget": "~2.3M/month shared · 2x", "contextWindow": 262144, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "mistral-large-latest", "displayName": "Mistral Large Latest (NavyAI)", "intelligenceRank": 14, "speedRank": 8, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 75000 }, "monthlyTokenBudget": "~2.3M/month shared · 2x", "contextWindow": 262144, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "sonar", "displayName": "Sonar (NavyAI)", "intelligenceRank": 22, "speedRank": 6, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 125000 }, "monthlyTokenBudget": "~3.8M/month shared · 1.2x", "contextWindow": 127072, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [] }, { "platform": "navy", "modelId": "sonar-pro", "displayName": "Sonar Pro (NavyAI)", "intelligenceRank": 13, "speedRank": 10, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 15000 }, "monthlyTokenBudget": "~450K/month shared · 10x", "contextWindow": 200000, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [] }, { "platform": "navy", "modelId": "sonar-reasoning-pro", "displayName": "Sonar Reasoning Pro (NavyAI)", "intelligenceRank": 10, "speedRank": 10, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 25000 }, "monthlyTokenBudget": "~750K/month shared · 6x", "contextWindow": 128000, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [] }, { "platform": "navy", "modelId": "sonar-deep-research", "displayName": "Sonar Deep Research (NavyAI)", "intelligenceRank": 10, "speedRank": 10, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 25000 }, "monthlyTokenBudget": "~750K/month shared · 6x", "contextWindow": 128000, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "navy", "modelId": "command-a-plus", "displayName": "Command A Plus (NavyAI)", "intelligenceRank": 10, "speedRank": 7, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 256000, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "navy", "modelId": "command-a", "displayName": "Command A (NavyAI)", "intelligenceRank": 10, "speedRank": 7, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 256000, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "navy", "modelId": "command-a-vision", "displayName": "Command A Vision (NavyAI)", "intelligenceRank": 10, "speedRank": 7, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 256000, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "navy", "modelId": "command-a-reasoning", "displayName": "Command A Reasoning (NavyAI)", "intelligenceRank": 10, "speedRank": 7, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 256000, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "navy", "modelId": "command-r-7b", "displayName": "Command R 7B (NavyAI)", "intelligenceRank": 22, "speedRank": 5, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 128000, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "navy", "modelId": "command-r", "displayName": "Command R (NavyAI)", "intelligenceRank": 22, "speedRank": 6, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 128000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "command-r-plus", "displayName": "Command R Plus (NavyAI)", "intelligenceRank": 17, "speedRank": 6, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 128000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "c4ai-aya-expanse-32b", "displayName": "C4ai Aya Expanse 32B (NavyAI)", "intelligenceRank": 22, "speedRank": 10, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": null, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "navy", "modelId": "c4ai-aya-vision-32b", "displayName": "C4ai Aya Vision 32B (NavyAI)", "intelligenceRank": 20, "speedRank": 10, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": null, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "navy", "modelId": "gpt-oss-20b", "displayName": "GPT Oss 20B (NavyAI)", "intelligenceRank": 14, "speedRank": 10, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "gpt-oss-120b", "displayName": "GPT Oss 120B (NavyAI)", "intelligenceRank": 6, "speedRank": 5, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "kimi-k2.7-code", "displayName": "Kimi K2.7 Code (NavyAI)", "intelligenceRank": 2, "speedRank": 10, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 50000 }, "monthlyTokenBudget": "~1.5M/month shared · 3x", "contextWindow": 262144, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "kimi-k2.6", "displayName": "Kimi K2.6 (NavyAI)", "intelligenceRank": 6, "speedRank": 8, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 53571 }, "monthlyTokenBudget": "~1.6M/month shared · 2.8x", "contextWindow": 262144, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "mimo-v2.5", "displayName": "MIMO V2.5 (NavyAI)", "intelligenceRank": 22, "speedRank": 6, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 125000 }, "monthlyTokenBudget": "~3.8M/month shared · 1.2x", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "mimo-v2.5-pro", "displayName": "MIMO V2.5 Pro (NavyAI)", "intelligenceRank": 10, "speedRank": 10, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 100000 }, "monthlyTokenBudget": "~3M/month shared · 1.5x", "contextWindow": 1048576, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "minimax-m3", "displayName": "Minimax M3 (NavyAI)", "intelligenceRank": 6, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 75000 }, "monthlyTokenBudget": "~2.3M/month shared · 2x", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "minimax-m2.7", "displayName": "Minimax M2.7 (NavyAI)", "intelligenceRank": 13, "speedRank": 5, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 93750 }, "monthlyTokenBudget": "~2.8M/month shared · 1.6x", "contextWindow": 204800, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "glm-5.2", "displayName": "GLM 5.2 (NavyAI)", "intelligenceRank": 4, "speedRank": 8, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 37500 }, "monthlyTokenBudget": "~1.1M/month shared · 4x", "contextWindow": 1048576, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "glm-5.1", "displayName": "GLM 5.1 (NavyAI)", "intelligenceRank": 6, "speedRank": 8, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 50000 }, "monthlyTokenBudget": "~1.5M/month shared · 3x", "contextWindow": 202752, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "qwen3.5-397b-a17b", "displayName": "Qwen3.5 397B A17b (NavyAI)", "intelligenceRank": 4, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 50000 }, "monthlyTokenBudget": "~1.5M/month shared · 3x", "contextWindow": 262144, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "nemotron-3-super", "displayName": "Nemotron 3 Super (NavyAI)", "intelligenceRank": 10, "speedRank": 7, "sizeLabel": "Frontier", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 1000000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "navy", "modelId": "hermes-4-70b", "displayName": "Hermes 4 70B (NavyAI)", "intelligenceRank": 17, "speedRank": 6, "sizeLabel": "Medium", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 150000 }, "monthlyTokenBudget": "~4.5M/month shared · 1x", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "navy", "modelId": "hermes-4-405b", "displayName": "Hermes 4 405B (NavyAI)", "intelligenceRank": 10, "speedRank": 10, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": null, "tpm": null, "tpd": 37500 }, "monthlyTokenBudget": "~1.1M/month shared · 4x", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "nvidia", "modelId": "openai/gpt-oss-120b", "displayName": "GPT-OSS 120B (NV)", "intelligenceRank": 6, "speedRank": 8, "sizeLabel": "Large", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ] }, { "platform": "nvidia", "modelId": "openai/gpt-oss-20b", "displayName": "GPT-OSS 20B (NV)", "intelligenceRank": 12, "speedRank": 6, "sizeLabel": "Small", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ] }, { "platform": "nvidia", "modelId": "z-ai/glm-5.2", "displayName": "GLM-5.2 (NV)", "intelligenceRank": 3, "speedRank": 7, "sizeLabel": "Frontier", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 200000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ] }, { "platform": "nvidia", "modelId": "nvidia/nemotron-3-ultra-550b-a55b", "displayName": "Nemotron-3 Ultra 550B (NV)", "intelligenceRank": 3, "speedRank": 8, "sizeLabel": "Frontier", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 1048576, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ] }, { "platform": "nvidia", "modelId": "nvidia/nvidia-nemotron-nano-9b-v2", "displayName": "Nemotron Nano 9B v2 (NV)", "intelligenceRank": 30, "speedRank": 5, "sizeLabel": "Small", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ] }, { "platform": "nvidia", "modelId": "meta/llama-3.2-90b-vision-instruct", "displayName": "Llama 3.2 90B Vision (NV)", "intelligenceRank": 18, "speedRank": 6, "sizeLabel": "Large", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 131072, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ] }, { "platform": "huggingface", "modelId": "deepseek-ai/DeepSeek-V3.2", "displayName": "DeepSeek V3.2 (HF)", "intelligenceRank": 5, "speedRank": 7, "sizeLabel": "Frontier", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "$0.10/mo credit", "contextWindow": 163840, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "hf-tiny-credit", "title": "Small $0.10/month routed credit", "body": "HuggingFace Inference Providers grants only ~$0.10/month of routed credit on the free tier (PRO is $2/month). Enough for light experimentation; exhausts quickly. Credits apply only to HF-routed requests.", "severity": "warning" } ] }, { "platform": "huggingface", "modelId": "deepseek-ai/DeepSeek-R1", "displayName": "DeepSeek R1 (HF)", "intelligenceRank": 5, "speedRank": 6, "sizeLabel": "Frontier", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "$0.10/mo credit", "contextWindow": 163840, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "hf-tiny-credit", "title": "Small $0.10/month routed credit", "body": "HuggingFace Inference Providers grants only ~$0.10/month of routed credit on the free tier (PRO is $2/month). Enough for light experimentation; exhausts quickly. Credits apply only to HF-routed requests.", "severity": "warning" } ] }, { "platform": "huggingface", "modelId": "Qwen/Qwen3-Coder-480B-A35B-Instruct", "displayName": "Qwen3-Coder 480B (HF)", "intelligenceRank": 3, "speedRank": 7, "sizeLabel": "Frontier", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "$0.10/mo credit", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "hf-tiny-credit", "title": "Small $0.10/month routed credit", "body": "HuggingFace Inference Providers grants only ~$0.10/month of routed credit on the free tier (PRO is $2/month). Enough for light experimentation; exhausts quickly. Credits apply only to HF-routed requests.", "severity": "warning" } ] }, { "platform": "huggingface", "modelId": "moonshotai/Kimi-K2.7-Code", "displayName": "Kimi K2.7 Code (HF)", "intelligenceRank": 3, "speedRank": 7, "sizeLabel": "Frontier", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "$0.10/mo credit", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "hf-tiny-credit", "title": "Small $0.10/month routed credit", "body": "HuggingFace Inference Providers grants only ~$0.10/month of routed credit on the free tier (PRO is $2/month). Enough for light experimentation; exhausts quickly. Credits apply only to HF-routed requests.", "severity": "warning" } ] }, { "platform": "huggingface", "modelId": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "displayName": "Llama 4 Maverick (HF)", "intelligenceRank": 10, "speedRank": 7, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "$0.10/mo credit", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [ { "slug": "hf-tiny-credit", "title": "Small $0.10/month routed credit", "body": "HuggingFace Inference Providers grants only ~$0.10/month of routed credit on the free tier (PRO is $2/month). Enough for light experimentation; exhausts quickly. Credits apply only to HF-routed requests.", "severity": "warning" } ] }, { "platform": "huggingface", "modelId": "MiniMaxAI/MiniMax-M3", "displayName": "MiniMax M3 (HF)", "intelligenceRank": 5, "speedRank": 7, "sizeLabel": "Frontier", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "$0.10/mo credit", "contextWindow": 1000192, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "hf-tiny-credit", "title": "Small $0.10/month routed credit", "body": "HuggingFace Inference Providers grants only ~$0.10/month of routed credit on the free tier (PRO is $2/month). Enough for light experimentation; exhausts quickly. Credits apply only to HF-routed requests.", "severity": "warning" } ] }, { "platform": "huggingface", "modelId": "zai-org/GLM-5.2", "displayName": "GLM-5.2 (HF)", "intelligenceRank": 3, "speedRank": 7, "sizeLabel": "Frontier", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "$0.10/mo credit", "contextWindow": 200000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "hf-tiny-credit", "title": "Small $0.10/month routed credit", "body": "HuggingFace Inference Providers grants only ~$0.10/month of routed credit on the free tier (PRO is $2/month). Enough for light experimentation; exhausts quickly. Credits apply only to HF-routed requests.", "severity": "warning" } ] }, { "platform": "huggingface", "modelId": "Qwen/Qwen3-VL-235B-A22B-Instruct", "displayName": "Qwen3-VL 235B (HF)", "intelligenceRank": 6, "speedRank": 6, "sizeLabel": "Frontier", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "$0.10/mo credit", "contextWindow": 131072, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [ { "slug": "hf-tiny-credit", "title": "Small $0.10/month routed credit", "body": "HuggingFace Inference Providers grants only ~$0.10/month of routed credit on the free tier (PRO is $2/month). Enough for light experimentation; exhausts quickly. Credits apply only to HF-routed requests.", "severity": "warning" } ] }, { "platform": "huggingface", "modelId": "openai/gpt-oss-120b", "displayName": "GPT-OSS 120B (HF)", "intelligenceRank": 6, "speedRank": 7, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "$0.10/mo credit", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "hf-tiny-credit", "title": "Small $0.10/month routed credit", "body": "HuggingFace Inference Providers grants only ~$0.10/month of routed credit on the free tier (PRO is $2/month). Enough for light experimentation; exhausts quickly. Credits apply only to HF-routed requests.", "severity": "warning" } ] }, { "platform": "huggingface", "modelId": "google/gemma-4-31B-it", "displayName": "Gemma 4 31B (HF)", "intelligenceRank": 20, "speedRank": 6, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "$0.10/mo credit", "contextWindow": 131072, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [ { "slug": "hf-tiny-credit", "title": "Small $0.10/month routed credit", "body": "HuggingFace Inference Providers grants only ~$0.10/month of routed credit on the free tier (PRO is $2/month). Enough for light experimentation; exhausts quickly. Credits apply only to HF-routed requests.", "severity": "warning" } ] }, { "platform": "cloudflare", "modelId": "@cf/qwen/qwen2.5-coder-32b-instruct", "displayName": "Qwen2.5 Coder 32B (CF)", "intelligenceRank": 15, "speedRank": 7, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~10-20M", "contextWindow": 32768, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ] }, { "platform": "cloudflare", "modelId": "@cf/qwen/qwq-32b", "displayName": "QwQ 32B (CF)", "intelligenceRank": 12, "speedRank": 6, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~10-20M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ] }, { "platform": "cloudflare", "modelId": "@cf/mistralai/mistral-small-3.1-24b-instruct", "displayName": "Mistral Small 3.1 24B (CF)", "intelligenceRank": 20, "speedRank": 7, "sizeLabel": "Medium", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~10-20M", "contextWindow": 131072, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ] }, { "platform": "cloudflare", "modelId": "@cf/meta/llama-3.2-11b-vision-instruct", "displayName": "Llama 3.2 11B Vision (CF)", "intelligenceRank": 30, "speedRank": 7, "sizeLabel": "Small", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~10-20M", "contextWindow": 131072, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ] }, { "platform": "cloudflare", "modelId": "@cf/meta/llama-3.1-8b-instruct-fp8", "displayName": "Llama 3.1 8B (CF)", "intelligenceRank": 40, "speedRank": 8, "sizeLabel": "8B", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~10-20M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ] }, { "platform": "cloudflare", "modelId": "@cf/meta/llama-3.2-3b-instruct", "displayName": "Llama 3.2 3B (CF)", "intelligenceRank": 60, "speedRank": 9, "sizeLabel": "3B", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~10-20M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ] }, { "platform": "cloudflare", "modelId": "@cf/aisingapore/gemma-sea-lion-v4-27b-it", "displayName": "Gemma SEA-LION v4 27B (CF)", "intelligenceRank": 30, "speedRank": 6, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~10-20M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ] }, { "platform": "cloudflare", "modelId": "@cf/meta/llama-guard-3-8b", "displayName": "Llama Guard 3 8B (CF)", "intelligenceRank": 200, "speedRank": 8, "sizeLabel": "8B", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~10-20M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ] }, { "platform": "groq", "modelId": "whisper-large-v3-turbo", "displayName": "Whisper Large v3 Turbo (Groq)", "intelligenceRank": 200, "speedRank": 1, "sizeLabel": "Audio", "limits": { "rpm": 30, "rpd": 1000, "tpm": 70000, "tpd": null }, "monthlyTokenBudget": "~6M", "contextWindow": null, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [], "modality": "audio", "mediaNote": "Speech-to-text" }, { "platform": "groq", "modelId": "whisper-large-v3", "displayName": "Whisper Large v3 (Groq)", "intelligenceRank": 200, "speedRank": 2, "sizeLabel": "Audio", "limits": { "rpm": 30, "rpd": 1000, "tpm": 70000, "tpd": null }, "monthlyTokenBudget": "~6M", "contextWindow": null, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [], "modality": "audio", "mediaNote": "Speech-to-text" }, { "platform": "cohere", "modelId": "command-a-translate-08-2025", "displayName": "Command A Translate (08-2025)", "intelligenceRank": 20, "speedRank": 8, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": 33, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~1-2M", "contextWindow": 256000, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "cohere", "modelId": "cohere-transcribe-03-2026", "displayName": "Cohere Transcribe (03-2026)", "intelligenceRank": 200, "speedRank": 5, "sizeLabel": "Audio", "limits": { "rpm": 20, "rpd": 33, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~1-2M", "contextWindow": null, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [], "modality": "audio", "mediaNote": "Speech-to-text" }, { "platform": "ollama", "modelId": "minimax-m3", "displayName": "MiniMax M3 (Ollama)", "intelligenceRank": 5, "speedRank": 8, "sizeLabel": "Frontier", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~5-10M", "contextWindow": 1000192, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "ollama", "modelId": "nemotron-3-super", "displayName": "Nemotron-3 Super (Ollama)", "intelligenceRank": 8, "speedRank": 8, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~5-10M", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [] }, { "platform": "sealion", "modelId": "aisingapore/Qwen-SEA-LION-v4.5-27B-IT", "displayName": "Qwen SEA-LION v4.5 27B (SEA-LION)", "intelligenceRank": 22, "speedRank": 7, "sizeLabel": "Medium", "limits": { "rpm": 10, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 10 RPM", "contextWindow": 32768, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "sealion", "modelId": "aisingapore/Qwen-SEA-LION-v4-32B-IT", "displayName": "Qwen SEA-LION v4 32B (SEA-LION)", "intelligenceRank": 24, "speedRank": 7, "sizeLabel": "Large", "limits": { "rpm": 10, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 10 RPM", "contextWindow": 32768, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "sealion", "modelId": "aisingapore/Gemma-SEA-LION-v4-27B-IT", "displayName": "Gemma SEA-LION v4 27B (SEA-LION)", "intelligenceRank": 26, "speedRank": 7, "sizeLabel": "Medium", "limits": { "rpm": 10, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 10 RPM", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "sealion", "modelId": "aisingapore/Llama-SEA-LION-v3-70B-IT", "displayName": "Llama SEA-LION v3 70B (SEA-LION)", "intelligenceRank": 20, "speedRank": 6, "sizeLabel": "Large", "limits": { "rpm": 10, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 10 RPM", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [] }, { "platform": "requesty", "modelId": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning", "displayName": "Nemotron 3 Nano Omni Reasoning (Requesty)", "intelligenceRank": 23, "speedRank": 9, "sizeLabel": "Medium", "limits": { "rpm": null, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 200 rpd", "contextWindow": 262144, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [], "premiumSince": "2026-07-19T18:44:22.000Z" }, { "platform": "cloudflare", "modelId": "@cf/meta/llama-3.2-1b-instruct", "displayName": "Llama 3.2 1B Instruct (CF)", "intelligenceRank": 70, "speedRank": 11, "sizeLabel": "1B", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~10-20M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ], "premiumSince": "2026-07-19T18:44:22.000Z" }, { "platform": "cloudflare", "modelId": "@cf/meta/llama-3.1-8b-instruct-fast", "displayName": "Llama 3.1 8B Instruct Fast (CF)", "intelligenceRank": 40, "speedRank": 11, "sizeLabel": "8B", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~10-20M", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ], "premiumSince": "2026-07-19T18:44:22.000Z" }, { "platform": "cloudflare", "modelId": "@cf/moondream/moondream3.1-9B-A2B", "displayName": "Moondream 3.1 9B A2B (CF)", "intelligenceRank": 35, "speedRank": 11, "sizeLabel": "9B", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~10-20M", "contextWindow": 32768, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info" } ], "premiumSince": "2026-07-19T18:44:22.000Z" }, { "platform": "google", "modelId": "gemini-3.6-flash", "displayName": "Gemini 3.6 Flash", "intelligenceRank": 2, "speedRank": 5, "sizeLabel": "Frontier", "limits": { "rpm": 10, "rpd": 20, "tpm": 250000, "tpd": null }, "monthlyTokenBudget": "~3M", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [ { "slug": "gemini-thinking-token-room", "title": "Thinking tokens use output cap", "body": "Gemini Flash can spend hidden thinking tokens inside maxOutputTokens. Avoid tiny max_tokens values or set thinkingBudget to 0 when provider support is available.", "severity": "info" } ], "premiumSince": "2026-07-22T09:39:29.000Z" }, { "platform": "google", "modelId": "gemini-3.5-flash-lite", "displayName": "Gemini 3.5 Flash Lite", "intelligenceRank": 12, "speedRank": 3, "sizeLabel": "Frontier", "limits": { "rpm": 15, "rpd": 20, "tpm": 250000, "tpd": null }, "monthlyTokenBudget": "~3M", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [ { "slug": "gemini-thinking-token-room", "title": "Thinking tokens use output cap", "body": "Gemini Flash can spend hidden thinking tokens inside maxOutputTokens. Avoid tiny max_tokens values or set thinkingBudget to 0 when provider support is available.", "severity": "info" } ], "premiumSince": "2026-07-22T09:39:29.000Z" }, { "platform": "google", "modelId": "gemini-robotics-er-1.6-preview", "displayName": "Gemini Robotics-ER 1.6 Preview", "intelligenceRank": 25, "speedRank": 5, "sizeLabel": "Medium", "limits": { "rpm": 10, "rpd": 20, "tpm": 250000, "tpd": null }, "monthlyTokenBudget": "~3M", "contextWindow": 131072, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [], "premiumSince": "2026-07-22T09:39:29.000Z" }, { "platform": "openrouter", "modelId": "poolside/laguna-s-2.1:free", "displayName": "Poolside Laguna S 2.1 (free)", "intelligenceRank": 12, "speedRank": 9, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "~6M", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "or-free-cap-account-wide", "title": "Daily :free cap is account-wide", "body": "OpenRouter’s :free daily cap (50/day, or 1000/day once you have ever bought $10 of credits) is shared across ALL :free models on the account, not per model. Per-row rpd values here are therefore optimistic; the router’s cooldown handling absorbs the shared 429s.", "severity": "info" } ], "premiumSince": "2026-07-22T09:39:29.000Z" }, { "platform": "kilo", "modelId": "poolside/laguna-s-2.1:free", "displayName": "Poolside Laguna S 2.1 (Kilo)", "intelligenceRank": 12, "speedRank": 8, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 200/hr per IP", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info" } ], "premiumSince": "2026-07-22T09:39:29.000Z" }, { "platform": "groq", "modelId": "meta-llama/llama-prompt-guard-2-22m", "displayName": "Llama Prompt Guard 2 22M (Groq)", "intelligenceRank": 200, "speedRank": 1, "sizeLabel": "Small", "limits": { "rpm": 30, "rpd": 14400, "tpm": 15000, "tpd": null }, "monthlyTokenBudget": "free · 14.4K rpd", "contextWindow": 512, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [], "premiumSince": "2026-07-22T09:39:29.000Z" }, { "platform": "groq", "modelId": "meta-llama/llama-prompt-guard-2-86m", "displayName": "Llama Prompt Guard 2 86M (Groq)", "intelligenceRank": 199, "speedRank": 1, "sizeLabel": "Small", "limits": { "rpm": 30, "rpd": 14400, "tpm": 15000, "tpd": null }, "monthlyTokenBudget": "free · 14.4K rpd", "contextWindow": 512, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [], "premiumSince": "2026-07-22T09:39:29.000Z" }, { "platform": "huggingface", "modelId": "zai-org/GLM-4.5", "displayName": "GLM-4.5 (HF)", "intelligenceRank": 7, "speedRank": 7, "sizeLabel": "Large", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "$0.10/mo credit", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "hf-tiny-credit", "title": "Small $0.10/month routed credit", "body": "HuggingFace Inference Providers grants only ~$0.10/month of routed credit on the free tier (PRO is $2/month). Enough for light experimentation; exhausts quickly. Credits apply only to HF-routed requests.", "severity": "warning" } ], "premiumSince": "2026-07-22T09:39:29.000Z" }, { "platform": "opencode", "modelId": "laguna-s-2.1-free", "displayName": "Laguna S 2.1 Free (OpenCode Zen)", "intelligenceRank": 12, "speedRank": 4, "sizeLabel": "Large", "limits": { "rpm": 20, "rpd": 200, "tpm": null, "tpd": null }, "monthlyTokenBudget": "promo (trial)", "contextWindow": 262144, "enabled": true, "supportsVision": false, "supportsTools": true, "quirks": [ { "slug": "zen-promo-roster", "title": "Limited-time promo, roster rotates", "body": "OpenCode Zen free models are explicitly limited-time promotional access (\"available for a limited time\" per the docs), not a recurring quota. The roster rotates: qwen3.6-plus and minimax-m3 promos already ended. Expect any row here to die without notice; prompts/outputs may be used for model improvement.", "severity": "warning" } ], "premiumSince": "2026-07-22T09:39:29.000Z" }, { "platform": "google", "modelId": "gemini-3.1-flash-tts-preview", "displayName": "Gemini 3.1 Flash TTS Preview", "intelligenceRank": 213, "speedRank": 213, "sizeLabel": "Flash", "limits": { "rpm": 3, "rpd": 15, "tpm": null, "tpd": null }, "monthlyTokenBudget": "", "contextWindow": null, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [], "modality": "audio", "mediaNote": "WAV/PCM · 30+ voices", "premiumSince": "2026-07-22T09:39:29.000Z" }, { "platform": "nara", "modelId": "agnes-2.0-flash", "displayName": "Agnes 2.0 Flash (NaraRouter)", "intelligenceRank": 17, "speedRank": 6, "sizeLabel": "Medium", "limits": { "rpm": 10, "rpd": null, "tpm": null, "tpd": 7000000 }, "monthlyTokenBudget": "free · 7M/day shared", "contextWindow": 262144, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [ { "slug": "nara-free-plan", "title": "Daily free quota", "body": "NaraRouter free access requires a no-card account key plus Telegram channel/link verification. The free plan is documented at 7M tokens/day and 10 req/min, with daily reset.", "severity": "info" } ], "premiumSince": "2026-07-23T22:58:14.000Z" }, { "platform": "nvidia", "modelId": "google/diffusiongemma-26b-a4b-it", "displayName": "DiffusionGemma 26B A4B IT (NV)", "intelligenceRank": 10, "speedRank": 6, "sizeLabel": "Medium", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 262144, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ], "premiumSince": "2026-07-23T22:58:14.000Z" }, { "platform": "nvidia", "modelId": "nvidia/ising-calibration-1.5-31b", "displayName": "Ising Calibration 1.5 31B (NV)", "intelligenceRank": 79, "speedRank": 7, "sizeLabel": "Medium", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 131072, "enabled": true, "supportsVision": true, "supportsTools": false, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ], "premiumSince": "2026-07-23T22:58:14.000Z" }, { "platform": "nvidia", "modelId": "nvidia/nemotron-3.5-content-safety", "displayName": "Nemotron 3.5 Content Safety (NV)", "intelligenceRank": 31, "speedRank": 9, "sizeLabel": "Small", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 131072, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ], "premiumSince": "2026-07-23T22:58:14.000Z" }, { "platform": "nvidia", "modelId": "thinkingmachines/inkling", "displayName": "Inkling (NV)", "intelligenceRank": 9, "speedRank": 8, "sizeLabel": "Frontier", "limits": { "rpm": 40, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "free · 40 RPM", "contextWindow": 1048576, "enabled": true, "supportsVision": true, "supportsTools": true, "quirks": [ { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info" } ], "premiumSince": "2026-07-23T22:58:14.000Z" }, { "platform": "aihorde", "modelId": "koboldcpp/Llama-3.2-3B", "displayName": "Llama 3.2 3B (AI Horde)", "intelligenceRank": 195, "speedRank": 206, "sizeLabel": "3B", "limits": { "rpm": null, "rpd": null, "tpm": null, "tpd": null }, "monthlyTokenBudget": "Kudos queue (no token cap)", "contextWindow": 4096, "enabled": true, "supportsVision": false, "supportsTools": false, "quirks": [ { "slug": "aihorde-anon-slow", "title": "Free volunteer queue, slow", "body": "AI Horde routes to volunteer-run workers through a priority queue, so latency is seconds to minutes, not the sub-second of hosted providers. The anonymous key 0000000000 runs at the lowest priority; register a free key at aihorde.net for higher priority. The provider uses a 120s timeout.", "severity": "warning" }, { "slug": "aihorde-no-tools", "title": "No tool calling", "body": "AI Horde's OpenAI-compatible proxy does not support function/tool calling. The provider drops tools, tool_choice and parallel_tool_calls so a tool-using request still completes as plain chat instead of failing.", "severity": "info" }, { "slug": "aihorde-usage-estimated", "title": "Usage is kudos; tokens estimated", "body": "The proxy returns usage as {\"kudos\": N} with no token counts, and rejects max_tokens below 16 and a non-array stop. The AIHordeProvider normalizes the request (floors max_tokens, wraps stop) and synthesizes prompt/completion token estimates so analytics and savings math aren't zero.", "severity": "info" }, { "slug": "aihorde-roster-rotates", "title": "Roster + context depend on online workers", "body": "Model availability changes as volunteer workers come and go, so a listed model can be temporarily unserved. The effective context window is set by the worker (often 4-8K), not the model's native maximum.", "severity": "info" }, { "slug": "aihorde-quality-uneven", "title": "Uneven quality", "body": "Output quality varies by worker and quantization, and some workers append template or instruction text after the answer. Best treated as free fallback capacity, not a primary model.", "severity": "warning" } ], "premiumSince": "2026-07-23T22:58:14.000Z" } ], "quirks": [ { "slug": "cloudflare-key-format", "title": "Key is account_id:token", "body": "Cloudflare Workers AI authenticates with a combined credential in the form \"account_id:token\", not a bare token.", "severity": "info", "targets": [ { "platform": "cloudflare", "modelGlob": null } ] }, { "slug": "keyless-anonymous", "title": "No API key required", "body": "Routes anonymously — the catalog ships a keyless sentinel row and calls work with no account or key.", "severity": "info", "targets": [ { "platform": "kilo", "modelGlob": null }, { "platform": "llm7", "modelGlob": null }, { "platform": "ovh", "modelGlob": null } ] }, { "slug": "nvidia-rate-limited", "title": "Recurring free, 40 RPM, eval-only ToS", "body": "NVIDIA NIM replaced its depleting trial credits with a recurring per-account rate limit (40 RPM default, varies by model), verified June 2026. The trial ToS still scopes usage to evaluation/prototyping, not production.", "severity": "info", "targets": [ { "platform": "nvidia", "modelGlob": null } ] }, { "slug": "or-free-cap-account-wide", "title": "Daily :free cap is account-wide", "body": "OpenRouter’s :free daily cap (50/day, or 1000/day once you have ever bought $10 of credits) is shared across ALL :free models on the account, not per model. Per-row rpd values here are therefore optimistic; the router’s cooldown handling absorbs the shared 429s.", "severity": "info", "targets": [ { "platform": "openrouter", "modelGlob": "*:free" } ] }, { "slug": "reasoning-token-room", "title": "Needs token room", "body": "Some free routes spend hidden reasoning tokens before visible output. Avoid tiny max_tokens values or requests can finish by length with empty content.", "severity": "info", "targets": [ { "platform": "openrouter", "modelGlob": "poolside/laguna-xs-2.1:free" }, { "platform": "kilo", "modelGlob": "poolside/laguna-xs-2.1:free" }, { "platform": "cloudflare", "modelGlob": "@cf/openai/gpt-oss-*" }, { "platform": "ovh", "modelGlob": "gpt-oss-20b" }, { "platform": "ovh", "modelGlob": "Qwen3.6-27B" }, { "platform": "requesty", "modelGlob": "poolside/laguna-m.1" } ] }, { "slug": "gemini-thinking-token-room", "title": "Thinking tokens use output cap", "body": "Gemini Flash can spend hidden thinking tokens inside maxOutputTokens. Avoid tiny max_tokens values or set thinkingBudget to 0 when provider support is available.", "severity": "info", "targets": [ { "platform": "google", "modelGlob": "gemini-3*flash*" } ] }, { "slug": "cohere-reasoning-output", "title": "May emit reasoning prose", "body": "Cohere Command A Reasoning can spend output budget explaining its thought process on terse prompts. Use a larger max_tokens cap or a stricter no-reasoning instruction when exact short answers are required.", "severity": "warning", "targets": [ { "platform": "cohere", "modelGlob": "command-a-reasoning-*" } ] }, { "slug": "or-ultra-hangs", "title": "OpenRouter ultra route hangs", "body": "nemotron-3-ultra (550B) on OpenRouter takes 180s+ even on trivial prompts (heavily congested), so its OR row is seeded disabled. Use the OpenCode Zen route instead.", "severity": "warning", "targets": [ { "platform": "openrouter", "modelGlob": "*nemotron-3-ultra*" } ] }, { "slug": "ovh-anon-trickle", "title": "Anonymous tier is 2 req/min", "body": "OVH AI Endpoints anonymous mode is documented at 2 req/min per IP per model (observed even stricter across models). The 400 req/min authenticated tier requires a Public Cloud project with a payment method, so the catalog ships the keyless path. Treat as a breadth/fallback tier, not a throughput tier.", "severity": "warning", "targets": [ { "platform": "ovh", "modelGlob": null } ] }, { "slug": "pollinations-degraded", "title": "Publishable key uses recurring shared capacity", "body": "Pollinations chat uses https://gen.pollinations.ai/v1 with a free API key. Free capacity is quest pollen earned through tasks on enter.pollinations.ai — it no longer auto-refills per IP; the legacy text host is retired.", "severity": "warning", "targets": [ { "platform": "pollinations", "modelGlob": null } ] }, { "slug": "zen-promo-roster", "title": "Limited-time promo, roster rotates", "body": "OpenCode Zen free models are explicitly limited-time promotional access (\"available for a limited time\" per the docs), not a recurring quota. The roster rotates: qwen3.6-plus and minimax-m3 promos already ended. Expect any row here to die without notice; prompts/outputs may be used for model improvement.", "severity": "warning", "targets": [ { "platform": "opencode", "modelGlob": null } ] }, { "slug": "zen-serves-ultra-fast", "title": "Zen serves the 550B fast", "body": "OpenCode Zen serves nemotron-3-ultra in ~2s with working tool calls where the OpenRouter route hangs — the live-verified path for this model.", "severity": "info", "targets": [ { "platform": "opencode", "modelGlob": "*nemotron-3-ultra*" } ] }, { "slug": "zhipu-shared-key", "title": "Works with existing Zhipu key", "body": "glm-4.6v-flash is listed Free on Z.AI and answers 200 with the existing bigmodel.cn key; vision and structured tool calls both live-verified.", "severity": "info", "targets": [ { "platform": "zhipu", "modelGlob": "*glm-4.6v*" } ] }, { "slug": "aihorde-anon-slow", "title": "Free volunteer queue, slow", "body": "AI Horde routes to volunteer-run workers through a priority queue, so latency is seconds to minutes, not the sub-second of hosted providers. The anonymous key 0000000000 runs at the lowest priority; register a free key at aihorde.net for higher priority. The provider uses a 120s timeout.", "severity": "warning", "targets": [ { "platform": "aihorde", "modelGlob": null } ] }, { "slug": "aihorde-no-tools", "title": "No tool calling", "body": "AI Horde's OpenAI-compatible proxy does not support function/tool calling. The provider drops tools, tool_choice and parallel_tool_calls so a tool-using request still completes as plain chat instead of failing.", "severity": "info", "targets": [ { "platform": "aihorde", "modelGlob": null } ] }, { "slug": "aihorde-usage-estimated", "title": "Usage is kudos; tokens estimated", "body": "The proxy returns usage as {\"kudos\": N} with no token counts, and rejects max_tokens below 16 and a non-array stop. The AIHordeProvider normalizes the request (floors max_tokens, wraps stop) and synthesizes prompt/completion token estimates so analytics and savings math aren't zero.", "severity": "info", "targets": [ { "platform": "aihorde", "modelGlob": null } ] }, { "slug": "aihorde-roster-rotates", "title": "Roster + context depend on online workers", "body": "Model availability changes as volunteer workers come and go, so a listed model can be temporarily unserved. The effective context window is set by the worker (often 4-8K), not the model's native maximum.", "severity": "info", "targets": [ { "platform": "aihorde", "modelGlob": null } ] }, { "slug": "aihorde-quality-uneven", "title": "Uneven quality", "body": "Output quality varies by worker and quantization, and some workers append template or instruction text after the answer. Best treated as free fallback capacity, not a primary model.", "severity": "warning", "targets": [ { "platform": "aihorde", "modelGlob": null } ] }, { "slug": "aihorde-behemoth-slow", "title": "123B has heavy queue (seeded disabled)", "body": "Behemoth-X-123B has deep queue depth (50s+ ETA even on trivial prompts). Seeded disabled; enable only if minute-scale latency is acceptable.", "severity": "warning", "targets": [ { "platform": "aihorde", "modelGlob": "*Behemoth-X-123B*" } ] }, { "slug": "kilo-auto-token-room", "title": "Auto route needs token room", "body": "Kilo auto/free can route to reasoning models that spend 100+ internal reasoning tokens. Avoid tiny max_tokens or a request can finish by length with empty content.", "severity": "info", "targets": [ { "platform": "kilo", "modelGlob": "kilo-auto/free" } ] }, { "slug": "safety-classifier-output", "title": "Safety classifier output", "body": "This route returns moderation/safety labels rather than normal assistant prose. Use it as a guardrail model, not a primary chat model.", "severity": "info", "targets": [ { "platform": "kilo", "modelGlob": "*content-safety*" }, { "platform": "ovh", "modelGlob": "Qwen3Guard-*" }, { "platform": "requesty", "modelGlob": "nvidia/nemotron-3.5-content-safety" } ] }, { "slug": "hy3-limited-free", "title": "Free route can change", "body": "Tencent Hy3 is currently free on these routes, but provider announcements describe some access as limited-time promotional availability. Monitor before aging into the monthly catalog.", "severity": "warning", "targets": [ { "platform": "requesty", "modelGlob": "novita/tencent/hy3" } ] }, { "slug": "nara-free-plan", "title": "Daily free quota", "body": "NaraRouter free access requires a no-card account key plus Telegram channel/link verification. The free plan is documented at 7M tokens/day and 10 req/min, with daily reset.", "severity": "info", "targets": [ { "platform": "nara", "modelGlob": null } ] }, { "slug": "ovh-qwen-think-trace", "title": "May emit thinking trace", "body": "OVH Qwen reasoning routes can include blocks or reasoning prose. Strip or suppress the thinking trace when a plain answer is required.", "severity": "warning", "targets": [ { "platform": "ovh", "modelGlob": "Qwen3*" } ] }, { "slug": "requesty-non-greedy-sampling", "title": "Requires non-greedy sampling", "body": "Requesty rejected this route when temperature was 0 with greedy sampling. Use a nonzero temperature or provider-compatible sampling parameters.", "severity": "info", "targets": [ { "platform": "requesty", "modelGlob": "mistral/leanstral-1-5" } ] }, { "slug": "hf-tiny-credit", "title": "Small $0.10/month routed credit", "body": "HuggingFace Inference Providers grants only ~$0.10/month of routed credit on the free tier (PRO is $2/month). Enough for light experimentation; exhausts quickly. Credits apply only to HF-routed requests.", "severity": "warning", "targets": [ { "platform": "huggingface", "modelGlob": null } ] }, { "slug": "github-free-quota", "title": "Daily GitHub Models free quota", "body": "GitHub Models free API quotas reset daily and are shared by rate-limit tier: low-tier models allow 15 RPM / 150 requests per day, high-tier models 10 RPM / 50 per day, and DeepSeek R1 1 RPM / 8 per day. Intended for experimentation, not production.", "severity": "info", "targets": [ { "platform": "github", "modelGlob": null } ] }, { "slug": "modelscope-aliyun-binding", "title": "Requires Alibaba Cloud (China) account binding", "body": "Free, but requires binding your ModelScope account to an Alibaba Cloud CHINA-site (cn) account with Chinese real-name verification — international (alibabacloud.com) accounts do not qualify (maintainer-confirmed). The API token mints without the binding, but every call then fails with `401 please bind your alibaba cloud account before use`.", "severity": "warning", "targets": [ { "platform": "modelscope", "modelGlob": null } ] }, { "slug": "bai-limited-time-free", "title": "Limited-time 0-credit promotion", "body": "B.AI currently bills DeepSeek V4 Flash at 0 Credits under a limited-time promotion that began August 17, 2026. B.AI says standard pricing resumes when the offer ends; re-check pricing before use after any provider notice.", "severity": "warning", "targets": [ { "platform": "bai", "modelGlob": "deepseek-v4-flash" } ] }, { "slug": "cn-real-name-verification", "title": "Requires Chinese real-name verification", "body": "Free, but the cloud account must pass Chinese real-name verification (实名认证) — normally a mainland ID or business licence — before the key will serve traffic. A key can usually be created without it, and then every call fails on auth or permission. Baidu Qianfan, Volcengine Ark and iFlytek Spark all sit behind this gate. LongCat is the exception: its platform accepts an email signup from outside mainland China.", "severity": "warning", "targets": [ { "platform": "qianfan", "modelGlob": null }, { "platform": "volcengine", "modelGlob": null }, { "platform": "xfyun", "modelGlob": null } ] } ], "embeddings": [ { "family": "gemini-embedding-001", "platform": "google", "modelId": "gemini-embedding-001", "displayName": "Gemini Embedding", "dimensions": 3072, "maxInputTokens": 2048, "priority": 1, "enabled": true, "quotaLabel": "100 rpm · 1K req/day", "premiumSince": "2026-07-19T18:44:22.000Z" }, { "family": "llama-nemotron-embed-vl-1b-v2", "platform": "nvidia", "modelId": "nvidia/llama-nemotron-embed-vl-1b-v2", "displayName": "Nemotron Embed VL 1B", "dimensions": 2048, "maxInputTokens": 8192, "priority": 1, "enabled": true, "quotaLabel": "~40 rpm", "premiumSince": "2026-07-19T18:44:22.000Z" }, { "family": "llama-nemotron-embed-vl-1b-v2", "platform": "openrouter", "modelId": "nvidia/llama-nemotron-embed-vl-1b-v2:free", "displayName": "Nemotron Embed VL 1B (OR free)", "dimensions": 2048, "maxInputTokens": 131072, "priority": 2, "enabled": true, "quotaLabel": "20 rpm · shared free daily cap", "premiumSince": "2026-07-19T18:44:22.000Z" }, { "family": "nemotron-3-embed-1b", "platform": "openrouter", "modelId": "nvidia/nemotron-3-embed-1b:free", "displayName": "Nemotron 3 Embed 1B (OR free)", "dimensions": 2048, "maxInputTokens": 32768, "priority": 1, "enabled": true, "quotaLabel": "20 rpm · shared free daily cap", "premiumSince": "2026-07-19T18:44:22.000Z" }, { "family": "llama-nemotron-embed-1b-v2", "platform": "nvidia", "modelId": "nvidia/llama-nemotron-embed-1b-v2", "displayName": "Nemotron Embed 1B", "dimensions": 2048, "maxInputTokens": 8192, "priority": 1, "enabled": true, "quotaLabel": "~40 rpm", "premiumSince": "2026-07-19T18:44:22.000Z" }, { "family": "nv-embedqa-e5-v5", "platform": "nvidia", "modelId": "nvidia/nv-embedqa-e5-v5", "displayName": "NV-EmbedQA E5 v5", "dimensions": 1024, "maxInputTokens": 512, "priority": 1, "enabled": true, "quotaLabel": "~40 rpm", "premiumSince": "2026-07-19T18:44:22.000Z" }, { "family": "sea-lion-e5-embedding-600m", "platform": "sealion", "modelId": "aisingapore/SEA-LION-E5-Embedding-600M", "displayName": "SEA-LION E5 Embedding 600M", "dimensions": 1024, "maxInputTokens": 8192, "priority": 1, "enabled": true, "quotaLabel": "free · 10 rpm", "premiumSince": "2026-07-19T18:44:22.000Z" }, { "family": "sea-lion-modernbert-embedding-300m", "platform": "sealion", "modelId": "aisingapore/SEA-LION-ModernBERT-Embedding-300M", "displayName": "SEA-LION ModernBERT Embedding 300M", "dimensions": 1024, "maxInputTokens": 8192, "priority": 1, "enabled": true, "quotaLabel": "free · 10 rpm", "premiumSince": "2026-07-19T18:44:22.000Z" }, { "family": "sea-lion-modernbert-embedding-600m", "platform": "sealion", "modelId": "aisingapore/SEA-LION-ModernBERT-Embedding-600M", "displayName": "SEA-LION ModernBERT Embedding 600M", "dimensions": 1024, "maxInputTokens": 8192, "priority": 1, "enabled": true, "quotaLabel": "free · 10 rpm", "premiumSince": "2026-07-19T18:44:22.000Z" }, { "family": "bge-m3", "platform": "cloudflare", "modelId": "@cf/baai/bge-m3", "displayName": "BGE-M3", "dimensions": 1024, "maxInputTokens": 8192, "priority": 1, "enabled": true, "quotaLabel": "10K neurons/day (shared)", "premiumSince": "2026-07-19T18:44:22.000Z" }, { "family": "bge-m3", "platform": "huggingface", "modelId": "BAAI/bge-m3", "displayName": "BGE-M3 (HF)", "dimensions": 1024, "maxInputTokens": 8192, "priority": 3, "enabled": true, "quotaLabel": "$0.10/mo credits", "premiumSince": "2026-07-19T18:44:22.000Z" }, { "family": "embeddinggemma-300m", "platform": "cloudflare", "modelId": "@cf/google/embeddinggemma-300m", "displayName": "EmbeddingGemma 300M", "dimensions": 768, "maxInputTokens": 2048, "priority": 1, "enabled": true, "quotaLabel": "10K neurons/day (shared)", "premiumSince": "2026-07-19T18:44:22.000Z" }, { "family": "qwen3-embedding-0.6b", "platform": "cloudflare", "modelId": "@cf/qwen/qwen3-embedding-0.6b", "displayName": "Qwen3 Embedding 0.6B", "dimensions": 1024, "maxInputTokens": 8192, "priority": 1, "enabled": true, "quotaLabel": "10K neurons/day (shared)", "premiumSince": "2026-07-19T18:44:22.000Z" }, { "family": "gemini-embedding-2", "platform": "google", "modelId": "gemini-embedding-2", "displayName": "Gemini Embedding 2", "dimensions": 3072, "maxInputTokens": 8192, "priority": 1, "enabled": true, "quotaLabel": "free-tier rate limited", "premiumSince": "2026-07-22T09:39:29.000Z" }, { "family": "bge-m3", "platform": "sealion", "modelId": "BAAI/bge-m3", "displayName": "BGE-M3 (SEA-LION)", "dimensions": 1024, "maxInputTokens": 8192, "priority": 2, "enabled": true, "quotaLabel": "free · 10 rpm", "premiumSince": "2026-07-22T09:39:29.000Z" }, { "family": "nemotron-3-embed-1b", "platform": "nvidia", "modelId": "nvidia/nemotron-3-embed-1b", "displayName": "Nemotron 3 Embed 1B", "dimensions": 2048, "maxInputTokens": 4096, "priority": 1, "enabled": true, "quotaLabel": "~40 rpm", "premiumSince": "2026-07-23T22:58:14.000Z" } ], "transcriptionModels": [ { "platform": "groq", "modelId": "whisper-large-v3-turbo", "displayName": "Whisper Large v3 Turbo (Groq)", "priority": 0, "enabled": true, "maxBytes": 26214400, "quotaLabel": "free tier - 25 MB uploads" }, { "platform": "groq", "modelId": "whisper-large-v3", "displayName": "Whisper Large v3 (Groq)", "priority": 1, "enabled": true, "maxBytes": 26214400, "quotaLabel": "free tier - 25 MB uploads" }, { "platform": "cloudflare", "modelId": "@cf/openai/whisper-large-v3-turbo", "displayName": "Whisper Large v3 Turbo (Cloudflare Workers AI)", "priority": 2, "enabled": true, "subtitleFormats": [ "vtt" ], "requestStyle": "json", "quotaLabel": "Workers AI free allocation" }, { "platform": "cloudflare", "modelId": "@cf/openai/whisper", "displayName": "Whisper (Cloudflare Workers AI)", "priority": 3, "enabled": true, "subtitleFormats": [ "vtt" ], "requestStyle": "binary", "quotaLabel": "Workers AI free allocation" } ], "videoModels": [] }