{ "generatedAt": "2026-10-16T16:13:25.378Z", "dateModified": "2026-10-16", "source": "OpenRouter public models API (https://openrouter.ai/api/v1/models)", "priceUnit": "USD per 1M tokens", "totalListedUpstream": 466, "models": [ { "id": "anthropic/claude-opus-5.5", "slug": "anthropic-claude-opus-5-5", "name": "Claude Opus 5.5", "provider": "anthropic", "segment": "frontier", "openWeights": false, "huggingFaceId": null, "inputPerM": 4, "outputPerM": 20, "tieredPricing": false, "contextTokens": 1000000, "inputModalities": [ "text", "image", "file" ], "outputModalities": [ "text" ], "intelligenceIndex": 57.6, "smartness": 5, "pagesAtOnce": 1500, "freeListing": false, "taskCosts": { "essay": 0.0282, "chat": 5.248, "pdf": 0.1476 }, "released": "2026-09-22", "description": "Claude Opus 5.5 is Anthropic's flagship model for demanding reasoning, coding, and long-horizon agentic work, succeeding Claude Opus 5. It is particularly strong at multi-step changes in large codebases, code...", "verdict": "The strongest all-round pick when quality matters more than cost, especially for big, complicated jobs like long coding projects.", "bestFor": [ "coding", "hard problems", "long documents" ], "watchOut": "One of the pricier options. Use a cheaper model for routine work.", "openrouterUrl": "https://openrouter.ai/anthropic/claude-opus-5.5" }, { "id": "anthropic/claude-sonnet-5.5", "slug": "anthropic-claude-sonnet-5-5", "name": "Claude Sonnet 5.5", "provider": "anthropic", "segment": "frontier", "openWeights": false, "huggingFaceId": null, "inputPerM": 2, "outputPerM": 10, "tieredPricing": false, "contextTokens": 1000000, "inputModalities": [ "text", "image", "file" ], "outputModalities": [ "text" ], "intelligenceIndex": 56, "smartness": 5, "pagesAtOnce": 1500, "freeListing": false, "taskCosts": { "essay": 0.0141, "chat": 2.624, "pdf": 0.0738 }, "released": "2026-09-28", "description": "Claude Sonnet 5.5 is Anthropic's Sonnet-class model for well-scoped everyday work, succeeding Claude Sonnet 5 as a direct upgrade. It is especially strong at building features, fixing bugs, and producing...", "verdict": "The sensible default: close to the top tier for writing, fixing code and everyday work, at about half the price of Claude Opus 5.5.", "bestFor": [ "writing", "coding", "everyday work" ], "watchOut": "Not quite the top tier for the very hardest jobs.", "openrouterUrl": "https://openrouter.ai/anthropic/claude-sonnet-5.5" }, { "id": "anthropic/claude-fable-5.1", "slug": "anthropic-claude-fable-5-1", "name": "Claude Fable 5.1", "provider": "anthropic", "segment": "frontier", "openWeights": false, "huggingFaceId": null, "inputPerM": 10, "outputPerM": 50, "tieredPricing": false, "contextTokens": 1000000, "inputModalities": [ "text", "image", "file" ], "outputModalities": [ "text" ], "intelligenceIndex": 53.4, "smartness": 5, "pagesAtOnce": 1500, "freeListing": false, "taskCosts": { "essay": 0.0705, "chat": 13.12, "pdf": 0.369 }, "released": "2026-09-01", "description": "Claude Fable 5.1 improves on Claude Fable 5 across the board, with the biggest gains in agentic coding, long-running agentic workflows, and knowledge work: long code refactors, front-end and visual...", "verdict": "A premium Claude model for long, multi-step jobs and knowledge work. It costs 2.5 times as much as Claude Opus 5.5.", "bestFor": [ "long multi-step jobs", "knowledge work" ], "watchOut": "Test whether the extra cost is worth it for your task before relying on it.", "openrouterUrl": "https://openrouter.ai/anthropic/claude-fable-5.1" }, { "id": "openai/gpt-6-astra", "slug": "openai-gpt-6-astra", "name": "GPT-6 Astra", "provider": "openai", "segment": "frontier", "openWeights": false, "huggingFaceId": null, "inputPerM": 10, "outputPerM": 50, "tieredPricing": true, "contextTokens": 1050000, "inputModalities": [ "file", "image", "text" ], "outputModalities": [ "text" ], "intelligenceIndex": 52.7, "smartness": 5, "pagesAtOnce": 1580, "freeListing": false, "taskCosts": { "essay": 0.0705, "chat": 13.12, "pdf": 0.369 }, "released": "2026-09-04", "description": "GPT-6 Astra is OpenAI's flagship model for demanding end-to-end work. It is suited for advanced analysis, software engineering, deep research, scientific work, and document creation, with particular strengths in long-horizon...", "verdict": "OpenAI's top model, built for deep research, programming and creating documents. Priced to match.", "bestFor": [ "deep research", "programming", "documents" ], "watchOut": "The most expensive tier on our list.", "openrouterUrl": "https://openrouter.ai/openai/gpt-6-astra" }, { "id": "openai/gpt-6.1-sol", "slug": "openai-gpt-6-1-sol", "name": "GPT-6.1 Sol", "provider": "openai", "segment": "frontier", "openWeights": false, "huggingFaceId": null, "inputPerM": 2, "outputPerM": 10, "tieredPricing": true, "contextTokens": 1050000, "inputModalities": [ "file", "image", "text" ], "outputModalities": [ "text" ], "intelligenceIndex": 51.8, "smartness": 5, "pagesAtOnce": 1580, "freeListing": false, "taskCosts": { "essay": 0.0141, "chat": 2.624, "pdf": 0.0738 }, "released": "2026-09-29", "description": "GPT-6.1 Sol is an upgrade to GPT-6 Sol from OpenAI, positioned below the flagship GPT-6 Astra in the GPT-6 series. It is suited for agentic coding, computer use, document-heavy professional...", "verdict": "Good value from OpenAI: strong at programming and working through long documents, at a mid-range price.", "bestFor": [ "programming", "documents" ], "watchOut": "Sits below GPT-6 Astra in OpenAI's own lineup. Very long inputs cost more.", "openrouterUrl": "https://openrouter.ai/openai/gpt-6.1-sol" }, { "id": "google/gemini-4-argon", "slug": "google-gemini-4-argon", "name": "Gemini 4 Argon", "provider": "google", "segment": "frontier", "openWeights": false, "huggingFaceId": null, "inputPerM": 2, "outputPerM": 10, "tieredPricing": false, "contextTokens": 1048576, "inputModalities": [ "audio", "file", "image", "text", "video" ], "outputModalities": [ "text" ], "intelligenceIndex": 52.6, "smartness": 5, "pagesAtOnce": 1570, "freeListing": false, "taskCosts": { "essay": 0.0141, "chat": 2.624, "pdf": 0.0738 }, "released": "2026-10-13", "description": "Gemini 4 Argon is Google's flagship reasoning model, built for complex software engineering, deep mathematical and scientific problem solving, and autonomous multi-step agents. Google trained it to keep context across...", "verdict": "Google's top model. It can read text, pictures, audio, video and files, at a mid-range price.", "bestFor": [ "pictures and video", "math and science", "hard problems" ], "watchOut": "Newer model, so there is less real-world feedback on it so far.", "openrouterUrl": "https://openrouter.ai/google/gemini-4-argon" }, { "id": "meta/muse-wonder", "slug": "meta-muse-wonder", "name": "Muse Wonder", "provider": "meta", "segment": "frontier", "openWeights": false, "huggingFaceId": null, "inputPerM": 4, "outputPerM": 20, "tieredPricing": false, "contextTokens": 1048576, "inputModalities": [ "text", "image", "video", "file", "audio" ], "outputModalities": [ "text" ], "intelligenceIndex": null, "smartness": null, "pagesAtOnce": 1570, "freeListing": false, "taskCosts": { "essay": 0.0282, "chat": 5.248, "pdf": 0.1476 }, "released": "2026-10-12", "description": "Muse Wonder is a frontier reasoning model from Meta, designed for complex personal and professional agentic tasks. It accepts text, images, video, audio, and PDF documents and returns text, with...", "verdict": "Meta's top model for complex personal and work tasks. It understands text, pictures, video, audio and PDFs.", "bestFor": [ "mixed media", "personal and work tasks" ], "watchOut": "Very new, and we don't have a smartness score for it yet.", "openrouterUrl": "https://openrouter.ai/meta/muse-wonder" }, { "id": "x-ai/grok-4.7", "slug": "x-ai-grok-4-7", "name": "Grok 4.7", "provider": "x-ai", "segment": "frontier", "openWeights": false, "huggingFaceId": null, "inputPerM": 2, "outputPerM": 6, "tieredPricing": true, "contextTokens": 500000, "inputModalities": [ "text", "image", "file" ], "outputModalities": [ "text" ], "intelligenceIndex": 46.4, "smartness": 4, "pagesAtOnce": 750, "freeListing": false, "taskCosts": { "essay": 0.0087, "chat": 2.544, "pdf": 0.071 }, "released": "2026-09-21", "description": "Grok 4.7 is SpaceXAI's flagship model for coding, agentic tasks, and knowledge work, succeeding Grok 4.6. It is particularly strong at long-running software engineering tasks, verifying its own work, and...", "verdict": "Grok's top model, aimed at programming and long, multi-step jobs, priced below most top models.", "bestFor": [ "programming", "long multi-step jobs" ], "watchOut": "Reads less at once than the million-token models, and costs more on very long inputs.", "openrouterUrl": "https://openrouter.ai/x-ai/grok-4.7" }, { "id": "mistralai/mistral-large-4-0", "slug": "mistralai-mistral-large-4-0", "name": "Mistral Large 4", "provider": "mistralai", "segment": "frontier", "openWeights": false, "huggingFaceId": null, "inputPerM": 0.68, "outputPerM": 2.09, "tieredPricing": false, "contextTokens": 1048576, "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "intelligenceIndex": 38.4, "smartness": 3, "pagesAtOnce": 1570, "freeListing": false, "taskCosts": { "essay": 0.003, "chat": 0.866, "pdf": 0.0242 }, "released": "2026-10-06", "description": "Mistral Large 4 is a frontier multimodal (text and image input) model from Mistral AI built for reasoning, coding, and agentic workloads. It offers a 1M-token context window with up...", "verdict": "A low-cost large model from Mistral. A good choice when budget is tight but you still want a big model.", "bestFor": [ "tight budgets", "general use" ], "watchOut": "Scores lower than the top models on the smartness measure we use.", "openrouterUrl": "https://openrouter.ai/mistralai/mistral-large-4-0" }, { "id": "cohere/command-a-plus", "slug": "cohere-command-a-plus", "name": "Command A+", "provider": "cohere", "segment": "frontier", "openWeights": false, "huggingFaceId": null, "inputPerM": 0.3, "outputPerM": 1.5, "tieredPricing": false, "contextTokens": 192000, "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "intelligenceIndex": null, "smartness": null, "pagesAtOnce": 290, "freeListing": false, "taskCosts": { "essay": 0.0021, "chat": 0.3936, "pdf": 0.0111 }, "released": "2026-09-22", "description": "Command A+ is Cohere's flagship model for enterprise agentic workflows. It accepts text and image inputs with a 192K context window, supports native tool calling with strict tool schemas, structured...", "verdict": "Built for business use, where answers must follow a strict format. Priced low.", "bestFor": [ "business tasks", "structured answers" ], "watchOut": "Reads fewer pages at once, and we don't have a smartness score for it.", "openrouterUrl": "https://openrouter.ai/cohere/command-a-plus" }, { "id": "anthropic/claude-haiku-5.5", "slug": "anthropic-claude-haiku-5-5", "name": "Claude Haiku 5.5", "provider": "anthropic", "segment": "small-fast", "openWeights": false, "huggingFaceId": null, "inputPerM": 0.1, "outputPerM": 0.5, "tieredPricing": true, "contextTokens": 1000000, "inputModalities": [ "text", "image", "file" ], "outputModalities": [ "text" ], "intelligenceIndex": 43.4, "smartness": 4, "pagesAtOnce": 1500, "freeListing": false, "taskCosts": { "essay": 0.0007, "chat": 0.1312, "pdf": 0.0037 }, "released": "2026-10-07", "description": "Claude Haiku 5.5 is Anthropic's small, fast model for high-volume, cost-sensitive work such as summarization, subagents, and browser use. It succeeds Claude Haiku 4.5 with stronger coding, computer use, and...", "verdict": "The cheap and fast Claude. Great for summaries, quick questions and anything you do a lot of.", "bestFor": [ "summaries", "quick questions", "high volume" ], "watchOut": "Costs more for very long inputs.", "openrouterUrl": "https://openrouter.ai/anthropic/claude-haiku-5.5" }, { "id": "openai/gpt-6-luna", "slug": "openai-gpt-6-luna", "name": "GPT-6 Luna", "provider": "openai", "segment": "small-fast", "openWeights": false, "huggingFaceId": null, "inputPerM": 0.1, "outputPerM": 0.5, "tieredPricing": true, "contextTokens": 1050000, "inputModalities": [ "file", "image", "text" ], "outputModalities": [ "text" ], "intelligenceIndex": 38.1, "smartness": 3, "pagesAtOnce": 1580, "freeListing": false, "taskCosts": { "essay": 0.0007, "chat": 0.1312, "pdf": 0.0037 }, "released": "2026-09-22", "description": "GPT-6 Luna is the fast, cost-efficient model in OpenAI's GPT-6 series, positioned below GPT-6 Sol. It is suited for high-volume and latency-sensitive workloads such as chat, classification, and lightweight agentic...", "verdict": "OpenAI's fast, cheap model for chat, sorting things into groups and light tasks.", "bestFor": [ "chat", "sorting and labeling", "light tasks" ], "watchOut": "Costs more for very long inputs.", "openrouterUrl": "https://openrouter.ai/openai/gpt-6-luna" }, { "id": "google/gemini-3.8-flash", "slug": "google-gemini-3-8-flash", "name": "Gemini 3.8 Flash", "provider": "google", "segment": "small-fast", "openWeights": false, "huggingFaceId": null, "inputPerM": 0.825, "outputPerM": 4.125, "tieredPricing": false, "contextTokens": 1048576, "inputModalities": [ "text", "image", "video", "file", "audio" ], "outputModalities": [ "text" ], "intelligenceIndex": 40.9, "smartness": 4, "pagesAtOnce": 1570, "freeListing": false, "taskCosts": { "essay": 0.0058, "chat": 1.0824, "pdf": 0.0304 }, "released": "2026-09-02", "description": "Gemini 3.8 Flash is Google's most intelligent Flash model with significant gains from 3.7 Flash across software engineering, agentic tasks, and multi-step reasoning.", "verdict": "Google's best fast model: a middle ground between cheap and capable. It understands audio and video too.", "bestFor": [ "cheap but capable", "audio and video" ], "watchOut": "Costs several times more than the cheapest options.", "openrouterUrl": "https://openrouter.ai/google/gemini-3.8-flash" }, { "id": "meta/muse-spark-1.3", "slug": "meta-muse-spark-1-3", "name": "Muse Spark 1.3", "provider": "meta", "segment": "small-fast", "openWeights": false, "huggingFaceId": null, "inputPerM": 1.25, "outputPerM": 4.25, "tieredPricing": false, "contextTokens": 1048576, "inputModalities": [ "text", "image", "video", "file" ], "outputModalities": [ "text" ], "intelligenceIndex": 48.1, "smartness": 4, "pagesAtOnce": 1570, "freeListing": false, "taskCosts": { "essay": 0.0061, "chat": 1.6, "pdf": 0.0447 }, "released": "2026-09-02", "description": "Muse Spark 1.3 is a multimodal reasoning model from Meta for long-running agentic, multi-agent, and coding workflows. It is designed to keep track of information across extended tasks, work through...", "verdict": "A mid-priced Meta model for long, multi-step jobs and programming. It scores close to the top models.", "bestFor": [ "programming", "long multi-step jobs" ], "watchOut": "Not as cheap as the fast models, and you can't download it.", "openrouterUrl": "https://openrouter.ai/meta/muse-spark-1.3" }, { "id": "deepseek/deepseek-v4.1-flash", "slug": "deepseek-deepseek-v4-1-flash", "name": "DeepSeek V4.1 Flash", "provider": "deepseek", "segment": "small-fast", "openWeights": true, "huggingFaceId": "deepseek-ai/DeepSeek-V4.1-Flash", "inputPerM": 0.15, "outputPerM": 0.6, "tieredPricing": true, "contextTokens": 1048576, "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "intelligenceIndex": 39.5, "smartness": 3, "pagesAtOnce": 1570, "freeListing": false, "taskCosts": { "essay": 0.0009, "chat": 0.1938, "pdf": 0.0054 }, "released": "2026-09-10", "description": "DeepSeek V4.1 Flash is a sparse mixture-of-experts model from DeepSeek, and the first built on the company's Causal Encoder-Decoder (CED) architecture. It activates 8B parameters on input and 16B on...", "verdict": "Very low price, can be downloaded and run yourself, and reads a very large amount at once. Hard to beat for big batches of work.", "bestFor": [ "big batches", "low cost" ], "watchOut": "The price on the listing changes by day of the week. Running it yourself takes serious computer hardware.", "openrouterUrl": "https://openrouter.ai/deepseek/deepseek-v4.1-flash" }, { "id": "qwen/qwen3.8-flash", "slug": "qwen-qwen3-8-flash", "name": "Qwen3.8 Flash", "provider": "qwen", "segment": "small-fast", "openWeights": true, "huggingFaceId": "Qwen/Qwen3.8-Flash-Next", "inputPerM": 0.15, "outputPerM": 0.47, "tieredPricing": false, "contextTokens": 1000000, "inputModalities": [ "text", "image", "video" ], "outputModalities": [ "text" ], "intelligenceIndex": null, "smartness": null, "pagesAtOnce": 1500, "freeListing": false, "taskCosts": { "essay": 0.0007, "chat": 0.1912, "pdf": 0.0053 }, "released": "2026-08-26", "description": "Qwen3.8 Flash is a multimodal reasoning model from Alibaba. It is suited for coding assistance, agentic workflows, visual understanding, document and codebase analysis, desktop interaction, chart analysis, and long-video analysis.", "verdict": "A cheap model that understands pictures and video, and can be downloaded and run yourself.", "bestFor": [ "pictures and video", "long documents" ], "watchOut": "We don't have a smartness score for it yet.", "openrouterUrl": "https://openrouter.ai/qwen/qwen3.8-flash" }, { "id": "xiaomi/mimo-v2.6-flash", "slug": "xiaomi-mimo-v2-6-flash", "name": "MiMo-V2.6-Flash", "provider": "xiaomi", "segment": "small-fast", "openWeights": true, "huggingFaceId": "XiaomiMiMo/MiMo-V2.6-Flash-RL", "inputPerM": 0.14, "outputPerM": 0.28, "tieredPricing": false, "contextTokens": 1050000, "inputModalities": [ "text", "image", "video", "audio" ], "outputModalities": [ "text" ], "intelligenceIndex": 37.9, "smartness": 3, "pagesAtOnce": 1580, "freeListing": false, "taskCosts": { "essay": 0.0004, "chat": 0.1753, "pdf": 0.0049 }, "released": "2026-09-21", "description": "MiMo-V2.6-Flash is an open-source foundation model developed by Xiaomi. Built on a Mixture-of-Experts architecture with 309B total parameters and 15B activated per token, it employs a hybrid attention mechanism for...", "verdict": "One of the cheapest models we track. It understands audio and video too, and can be downloaded and run yourself.", "bestFor": [ "very low cost", "audio and video" ], "watchOut": "Scores lower than the pricier models.", "openrouterUrl": "https://openrouter.ai/xiaomi/mimo-v2.6-flash" }, { "id": "meta/muse-spark-open", "slug": "meta-muse-spark-open", "name": "Muse Spark Open", "provider": "meta", "segment": "open-weights", "openWeights": true, "huggingFaceId": "meta-models/Muse-Spark-Open-816B-A42B", "inputPerM": 1.25, "outputPerM": 4.25, "tieredPricing": false, "contextTokens": 1048576, "inputModalities": [ "text", "image", "video", "file", "audio" ], "outputModalities": [ "text" ], "intelligenceIndex": null, "smartness": null, "pagesAtOnce": 1570, "freeListing": false, "taskCosts": { "essay": 0.0061, "chat": 1.6, "pdf": 0.0447 }, "released": "2026-10-15", "description": "Muse Spark Open is a coding-optimized open-weight model. Trained specifically for agentic coding workflows, it delivers higher first-attempt accuracy, reliable tool calling, and deep codebase understanding across the repos developers...", "verdict": "A downloadable Meta model tuned for programming. The one to watch if you want to run your own coding helper.", "bestFor": [ "running your own coding helper" ], "watchOut": "Brand new (listed Oct 15, 2026) with no smartness score yet. Check the computer requirements before planning to run it yourself.", "openrouterUrl": "https://openrouter.ai/meta/muse-spark-open" }, { "id": "xiaomi/mimo-v2.6-pro", "slug": "xiaomi-mimo-v2-6-pro", "name": "MiMo-V2.6-Pro", "provider": "xiaomi", "segment": "open-weights", "openWeights": true, "huggingFaceId": "XiaomiMiMo/MiMo-V2.6-Pro-RL", "inputPerM": 0.435, "outputPerM": 0.87, "tieredPricing": false, "contextTokens": 1050000, "inputModalities": [ "text", "image", "video", "audio" ], "outputModalities": [ "text" ], "intelligenceIndex": 46.3, "smartness": 4, "pagesAtOnce": 1580, "freeListing": false, "taskCosts": { "essay": 0.0013, "chat": 0.5446, "pdf": 0.0151 }, "released": "2026-09-21", "description": "MiMo-V2.6-Pro is the flagship foundation model developed by Xiaomi. Built at a scale of over 1T parameters, it is designed to push the ceiling of capability for the most demanding...", "verdict": "The smartest downloadable model we track, at a very low price.", "bestFor": [ "good value", "general use", "coding" ], "watchOut": "Too big to run on a normal computer, so most people would use it through a paid service.", "openrouterUrl": "https://openrouter.ai/xiaomi/mimo-v2.6-pro" }, { "id": "moonshotai/kimi-k3", "slug": "moonshotai-kimi-k3", "name": "Kimi K3", "provider": "moonshotai", "segment": "open-weights", "openWeights": true, "huggingFaceId": "moonshotai/Kimi-K3", "inputPerM": 0.304, "outputPerM": 11.936, "tieredPricing": false, "contextTokens": 1048576, "inputModalities": [ "text", "image", "video" ], "outputModalities": [ "text" ], "intelligenceIndex": 43.6, "smartness": 4, "pagesAtOnce": 1570, "freeListing": false, "taskCosts": { "essay": 0.0162, "chat": 0.6072, "pdf": 0.0185 }, "released": "2026-07-16", "description": "Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI. It is suited for complex coding, knowledge work, and long-horizon agentic workflows, and is particularly strong at...", "verdict": "A very large downloadable model for complicated programming and long jobs.", "bestFor": [ "complicated coding", "long jobs" ], "watchOut": "Replies cost a lot for a downloadable model on the current listing. Check the price before committing.", "openrouterUrl": "https://openrouter.ai/moonshotai/kimi-k3" }, { "id": "z-ai/glm-5.3", "slug": "z-ai-glm-5-3", "name": "GLM 5.3", "provider": "z-ai", "segment": "open-weights", "openWeights": true, "huggingFaceId": "zai-org/GLM-5.3", "inputPerM": 0.079, "outputPerM": 4.8, "tieredPricing": false, "contextTokens": 1048576, "inputModalities": [ "text" ], "outputModalities": [ "text" ], "intelligenceIndex": 44.8, "smartness": 4, "pagesAtOnce": 1570, "freeListing": false, "taskCosts": { "essay": 0.0065, "chat": 0.1917, "pdf": 0.006 }, "released": "2026-08-18", "description": "GLM-5.3 is a large-scale reasoning model from Z.ai, built for complex software engineering and long-horizon agent tasks. It supports text input and output with a 1M-token context window, and improves...", "verdict": "A downloadable model for software work and long jobs, with very cheap input.", "bestFor": [ "software work", "long jobs" ], "watchOut": "Text only. Replies cost far more than what you send in.", "openrouterUrl": "https://openrouter.ai/z-ai/glm-5.3" }, { "id": "qwen/qwen3.8-27b", "slug": "qwen-qwen3-8-27b", "name": "Qwen3.8 27B", "provider": "qwen", "segment": "open-weights", "openWeights": true, "huggingFaceId": "Qwen/Qwen3.8-27B", "inputPerM": 0.24, "outputPerM": 2.2, "tieredPricing": false, "contextTokens": 1000000, "inputModalities": [ "text", "image", "video" ], "outputModalities": [ "text" ], "intelligenceIndex": 33.7, "smartness": 3, "pagesAtOnce": 1500, "freeListing": false, "taskCosts": { "essay": 0.003, "chat": 0.3349, "pdf": 0.0096 }, "released": "2026-08-14", "description": "Qwen3.8 27B is an open-weight dense vision-language model from Qwen. It is suited for coding, professional workflows, research, multimodal interaction, and long-running agent tasks, with flexible thinking that can be...", "verdict": "A mid-size downloadable model that understands pictures and video, and is realistic to run on your own hardware.", "bestFor": [ "running it yourself", "pictures and video" ], "watchOut": "Scores below the big models.", "openrouterUrl": "https://openrouter.ai/qwen/qwen3.8-27b" }, { "id": "nvidia/nemotron-3.5-super-vl", "slug": "nvidia-nemotron-3-5-super-vl", "name": "Nemotron 3.5 Super VL", "provider": "nvidia", "segment": "open-weights", "openWeights": true, "huggingFaceId": "nvidia/NVIDIA-Nemotron-3.5-Super-VL", "inputPerM": 0.15, "outputPerM": 0.5, "tieredPricing": false, "contextTokens": 1048576, "inputModalities": [ "text", "image", "video" ], "outputModalities": [ "text" ], "intelligenceIndex": 26.5, "smartness": 2, "pagesAtOnce": 1570, "freeListing": true, "taskCosts": { "essay": 0.0007, "chat": 0.1918, "pdf": 0.0054 }, "released": "2026-10-15", "description": "NVIDIA Nemotron 3.5 Super VL is an open multimodal mixture-of-experts model from NVIDIA, with 12B active parameters out of 120B total. It takes text, images, documents, GUI screenshots, and video...", "verdict": "A downloadable model that reads documents, screenshots and video at low cost. A free listing exists on OpenRouter.", "bestFor": [ "documents and screenshots", "video" ], "watchOut": "The lowest smartness score of the large models here.", "openrouterUrl": "https://openrouter.ai/nvidia/nemotron-3.5-super-vl" }, { "id": "openai/gpt-oss-120b", "slug": "openai-gpt-oss-120b", "name": "gpt-oss-120b", "provider": "openai", "segment": "open-weights", "openWeights": true, "huggingFaceId": "openai/gpt-oss-120b", "inputPerM": 0.15, "outputPerM": 0.6, "tieredPricing": false, "contextTokens": 131072, "inputModalities": [ "text" ], "outputModalities": [ "text" ], "intelligenceIndex": 11.6, "smartness": 1, "pagesAtOnce": 200, "freeListing": false, "taskCosts": { "essay": 0.0009, "chat": 0.1938, "pdf": 0.0054 }, "released": "2025-08-05", "description": "gpt-oss-120b is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases. It activates 5.1B parameters per forward pass and is optimized...", "verdict": "OpenAI's older downloadable model. Still a cheap, well-known starting point for running your own.", "bestFor": [ "running it yourself", "tight budgets" ], "watchOut": "Released August 2025 and well behind current models. Reads fewer pages at once.", "openrouterUrl": "https://openrouter.ai/openai/gpt-oss-120b" }, { "id": "google/gemma-4-31b-it", "slug": "google-gemma-4-31b-it", "name": "Gemma 4 31B", "provider": "google", "segment": "open-weights", "openWeights": true, "huggingFaceId": "google/gemma-4-31B-it", "inputPerM": 0.2, "outputPerM": 0.4, "tieredPricing": false, "contextTokens": 262144, "inputModalities": [ "image", "text", "video" ], "outputModalities": [ "text" ], "intelligenceIndex": 14.7, "smartness": 1, "pagesAtOnce": 390, "freeListing": true, "taskCosts": { "essay": 0.0006, "chat": 0.2504, "pdf": 0.007 }, "released": "2026-04-02", "description": "Gemma 4 31B Instruct is Google DeepMind's 30.7B dense multimodal model supporting text and image input with text output. Features a 256K token context window, configurable thinking/reasoning mode, native function...", "verdict": "Google's downloadable model, small enough to run yourself. It understands pictures and video. A free listing exists on OpenRouter.", "bestFor": [ "running it yourself", "small projects" ], "watchOut": "Well below the top models in quality.", "openrouterUrl": "https://openrouter.ai/google/gemma-4-31b-it" }, { "id": "deepseek/deepseek-v4-pro-0813", "slug": "deepseek-deepseek-v4-pro-0813", "name": "DeepSeek V4 Pro 0813", "provider": "deepseek", "segment": "open-weights", "openWeights": true, "huggingFaceId": "deepseek-ai/DeepSeek-V4-Pro-0813", "inputPerM": 0.66, "outputPerM": 1.98, "tieredPricing": true, "contextTokens": 1048576, "inputModalities": [ "text" ], "outputModalities": [ "text" ], "intelligenceIndex": 36, "smartness": 3, "pagesAtOnce": 1570, "freeListing": false, "taskCosts": { "essay": 0.0029, "chat": 0.8395, "pdf": 0.0234 }, "released": "2026-08-12", "description": "DeepSeek V4 Pro 0813 is a large-scale mixture-of-experts model from DeepSeek. This is the GA release of DeepSeek V4 Pro.", "verdict": "DeepSeek's larger downloadable model at a low price. A cheap step up from DeepSeek V4.1 Flash.", "bestFor": [ "low-cost quality writing and analysis" ], "watchOut": "Text only.", "openrouterUrl": "https://openrouter.ai/deepseek/deepseek-v4-pro-0813" }, { "id": "liquid/lfm-2.5-24b-a2b", "slug": "liquid-lfm-2-5-24b-a2b", "name": "LFM2.5-24B-A2B", "provider": "liquid", "segment": "open-weights", "openWeights": true, "huggingFaceId": "LiquidAI/LFM2.5-24B-A2B", "inputPerM": 0.05, "outputPerM": 0.7, "tieredPricing": false, "contextTokens": 128000, "inputModalities": [ "text" ], "outputModalities": [ "text" ], "intelligenceIndex": null, "smartness": null, "pagesAtOnce": 190, "freeListing": false, "taskCosts": { "essay": 0.001, "chat": 0.0746, "pdf": 0.0022 }, "released": "2026-10-13", "description": "LFM2.5-24B-A2B is a mixture-of-experts reasoning model from Liquid AI with 24B total and 2.45B active parameters. It always reasons before answering and is built for agents that call tools. Liquid...", "verdict": "A small downloadable model built to run on modest hardware.", "bestFor": [ "running on modest hardware" ], "watchOut": "Reads fewer pages at once, and has no smartness score yet.", "openrouterUrl": "https://openrouter.ai/liquid/lfm-2.5-24b-a2b" }, { "id": "inference-net/schematron-v2-turbo", "slug": "inference-net-schematron-v2-turbo", "name": "Schematron V2 Turbo", "provider": "inference-net", "segment": "specialist", "openWeights": true, "huggingFaceId": "inference-net/schematron-v2-granite-4.0-h-micro", "inputPerM": 0.03, "outputPerM": 0.15, "tieredPricing": false, "contextTokens": 128000, "inputModalities": [ "text" ], "outputModalities": [ "text" ], "intelligenceIndex": null, "smartness": null, "pagesAtOnce": 190, "freeListing": false, "taskCosts": { "essay": 0.0002, "chat": 0.0394, "pdf": 0.0011 }, "released": "2026-09-12", "description": "Schematron V2 Turbo is a 3B-parameter HTML-to-JSON extraction model from Inference.net. It prioritizes throughput for high-volume extraction workloads. Extraction instructions must be supplied through a JSON schema in response_format rather...", "verdict": "A tiny model that does one thing: turn a web page into a neat table of data, very cheaply.", "bestFor": [ "pulling data from web pages" ], "watchOut": "Not a general chat model.", "openrouterUrl": "https://openrouter.ai/inference-net/schematron-v2-turbo" }, { "id": "moondream/moondream3.1-9b-a2b", "slug": "moondream-moondream3-1-9b-a2b", "name": "Moondream 3.1", "provider": "moondream", "segment": "specialist", "openWeights": true, "huggingFaceId": "moondream/moondream3.1-9B-A2B", "inputPerM": 0.3, "outputPerM": 2.5, "tieredPricing": false, "contextTokens": 32000, "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "intelligenceIndex": null, "smartness": null, "pagesAtOnce": 50, "freeListing": false, "taskCosts": { "essay": 0.0035, "chat": 0.4136, "pdf": 0.0118 }, "released": "2026-10-15", "description": "Moondream 3.1 is a vision-language model from Moondream, the successor to Moondream 3 Preview, with 9B total parameters and 2B active (mixture-of-experts). It handles visual question answering, captioning, OCR, and...", "verdict": "A small model for answering questions about images, describing them and reading text in them.", "bestFor": [ "describing images", "reading text in photos" ], "watchOut": "Reads very little at once and gives short answers.", "openrouterUrl": "https://openrouter.ai/moondream/moondream3.1-9b-a2b" }, { "id": "google/gemini-nano-banana-2.1", "slug": "google-gemini-nano-banana-2-1", "name": "Nano Banana 2.1", "provider": "google", "segment": "specialist", "openWeights": false, "huggingFaceId": null, "inputPerM": 1.5, "outputPerM": 7.5, "tieredPricing": false, "contextTokens": 65536, "inputModalities": [ "image", "text" ], "outputModalities": [ "image", "text" ], "intelligenceIndex": null, "smartness": null, "pagesAtOnce": 100, "freeListing": false, "taskCosts": { "essay": 0.0106, "chat": 1.968, "pdf": 0.0554 }, "released": "2026-10-06", "description": "Nano Banana 2.1 (Gemini Nano Banana 2.1) is Google's image generation and editing model on the Flash tier, succeeding Nano Banana 2 and Nano Banana Pro. It improves product recontextualization,...", "verdict": "Google's model for creating and editing images, for product shots and photo edits.", "bestFor": [ "making images", "editing photos" ], "watchOut": "Image output is billed separately from text, so the costs on this site don't cover it.", "openrouterUrl": "https://openrouter.ai/google/gemini-nano-banana-2.1" }, { "id": "tencent/hy-mt2-7b", "slug": "tencent-hy-mt2-7b", "name": "Hy-MT2-7B", "provider": "tencent", "segment": "specialist", "openWeights": true, "huggingFaceId": "tencent/Hy-MT2-7B", "inputPerM": 0.074, "outputPerM": 0.295, "tieredPricing": false, "contextTokens": 8192, "inputModalities": [ "text" ], "outputModalities": [ "text" ], "intelligenceIndex": null, "smartness": null, "pagesAtOnce": 10, "freeListing": false, "taskCosts": { "essay": 0.0004, "chat": 0.0956, "pdf": 0.0027 }, "released": "2026-08-19", "description": "Hy-MT2-7B is a 7B-parameter translation model from Tencent. It supports 33 language pairs and five Chinese dialect and minority-language pairs, with workflows for structured, delimiter-based, contextual, glossary-based, and style-guided translation.", "verdict": "A model built only for translation, with options for glossaries and writing style. Covers 33 language pairs plus some Chinese dialects.", "bestFor": [ "translation" ], "watchOut": "Short inputs only, and translation only.", "openrouterUrl": "https://openrouter.ai/tencent/hy-mt2-7b" }, { "id": "perplexity/sonar-pro", "slug": "perplexity-sonar-pro", "name": "Sonar Pro", "provider": "perplexity", "segment": "specialist", "openWeights": false, "huggingFaceId": null, "inputPerM": 3, "outputPerM": 15, "tieredPricing": false, "contextTokens": 200000, "inputModalities": [ "text", "image" ], "outputModalities": [ "text" ], "intelligenceIndex": 7.6, "smartness": 1, "pagesAtOnce": 300, "freeListing": false, "taskCosts": { "essay": 0.0212, "chat": 3.936, "pdf": 0.1107 }, "released": "2025-03-07", "description": "Note: Sonar Pro pricing includes Perplexity search pricing. See [details here](https://docs.perplexity.ai/guides/pricing#detailed-pricing-breakdown-for-sonar-reasoning-pro-and-sonar-pro) For enterprises seeking more advanced capabilities, the Sonar Pro API can handle in-depth, multi-step queries with added extensibility, like...", "verdict": "Answers questions by searching the web and showing its sources. A research helper rather than a general chatbot.", "bestFor": [ "finding answers with sources" ], "watchOut": "The price includes Perplexity's search fees. An older model (2025).", "openrouterUrl": "https://openrouter.ai/perplexity/sonar-pro" } ] }