{"status":"success","total":17,"updatedAt":"2026-09-02T17:14:34.570Z","metadata":{"registry":"ModelRegistry","documentation":"https://modelregistry.tirup.in","license":"Open Data / MIT","description":"Open frontier AI model registry tracking primary flagships and research checkpoints across all premier labs."},"companies":[{"id":"anthropic","name":"Anthropic","latestFlagship":"Claude Fable 5.1","latestReasoning":"Claude Mythos 5.1 (Restricted) / Claude Fable 5.1"},{"id":"openai","name":"OpenAI","latestFlagship":"GPT-5.6 Sol / Terra Pro","latestReasoning":"OpenAI Astra (Cybersecurity) / GPT-5.6 Sol"},{"id":"google","name":"Google DeepMind","latestFlagship":"Gemini 3.8 Flash","latestReasoning":"Gemini 3.8 Flash / Gemini 3.1 Pro"},{"id":"deepseek","name":"DeepSeek","latestFlagship":"DeepSeek V4-Pro (0813)","latestReasoning":"DeepSeek V4-Pro (Thinking Mode)"},{"id":"xai","name":"xAI","latestFlagship":"Grok 4.6","latestReasoning":"Grok 4.6 (Think Mode)"},{"id":"qwen","name":"Alibaba Cloud (Qwen)","latestFlagship":"Qwen3.8 Max / 2.4T A95B","latestReasoning":"Qwen3.8 Flash (Multimodal Reasoning)"},{"id":"meta","name":"Meta AI","latestFlagship":"Llama 4 Maverick (128E)","latestReasoning":"Llama 4 Scout (16E)"},{"id":"mistral","name":"Mistral AI","latestFlagship":"Mistral Medium 3.5","latestReasoning":"Mistral Small 4"},{"id":"cohere","name":"Cohere","latestFlagship":"Command A (111B)","latestReasoning":"North Mini Code"}],"models":[{"id":"claude-fable-5-1","companyId":"anthropic","companyName":"Anthropic","name":"Claude Fable 5.1","version":"5.1","releaseDate":"2026-09-01","isCompanyFlagship":true,"isLatestCheckpoint":true,"statusBadge":"FRONTIER FLAGSHIP #1","category":"reasoning","categoryLabel":"Adaptive Reasoning","contextWindow":"1,000,000 tokens","contextWindowTokens":1000000,"maxOutputTokens":"128,000 tokens","parameters":"Frontier MoE (Adaptive Thinking)","openWeights":false,"license":"Proprietary API / Cloud Foundry","pricing":{"input":10,"output":50},"highlight":"Released Sept 1, 2026; Anthropic's reigning flagship for agentic coding and knowledge work. Features a 75% reduction in cache read costs ($0.25/M tokens) and 52.6% on Terminal-Bench-Science.","modalities":["Text","Vision","Code"],"benchmarks":{"terminalBench":"52.6%","sweBench":"78.4%","gpqa":"74.8%"},"links":{"announcement":"https://www.anthropic.com/news/claude-fable-5-1","playground":"https://claude.ai","apiDocs":"https://docs.anthropic.com/claude/reference"}},{"id":"claude-opus-5","companyId":"anthropic","companyName":"Anthropic","name":"Claude Opus 5","version":"5.0","releaseDate":"2026-07-15","isCompanyFlagship":false,"isLatestCheckpoint":false,"statusBadge":"HEAVYWEIGHT AGENTIC","category":"flagship","categoryLabel":"Deep Research","contextWindow":"1,000,000 tokens","contextWindowTokens":1000000,"maxOutputTokens":"65,536 tokens","parameters":"Dense Frontier","openWeights":false,"license":"Proprietary API","pricing":{"input":15,"output":75},"highlight":"Deep multi-hour research, complex code synthesis, and long-horizon workflow orchestration.","modalities":["Text","Vision","Code"],"benchmarks":{"sweBench":"74.2%","gpqa":"72.9%"},"links":{"playground":"https://claude.ai","apiDocs":"https://docs.anthropic.com"}},{"id":"gpt-5-6-sol","companyId":"openai","companyName":"OpenAI","name":"GPT-5.6 Sol","version":"5.6","releaseDate":"2026-08-05","isCompanyFlagship":true,"isLatestCheckpoint":false,"statusBadge":"PRIMARY FLAGSHIP","category":"flagship","categoryLabel":"Frontier Foundation","contextWindow":"1,050,000 tokens","contextWindowTokens":1050000,"maxOutputTokens":"65,536 tokens","parameters":"Unified Reasoning Architecture","openWeights":false,"license":"Proprietary API / ChatGPT Pro","pricing":{"input":3,"output":12},"highlight":"OpenAI's primary general-purpose flagship model. Combines native fast execution, multi-agent orchestration, and 1.05M context for production software development.","modalities":["Text","Vision","Audio","Code"],"benchmarks":{"mmluPro":"86.4%","sweBench":"76.8%"},"links":{"announcement":"https://openai.com/index/gpt-5-6/","playground":"https://chatgpt.com","apiDocs":"https://platform.openai.com/docs"}},{"id":"openai-astra","companyId":"openai","companyName":"OpenAI","name":"OpenAI Astra","version":"1.0","releaseDate":"2026-09-01","isCompanyFlagship":false,"isLatestCheckpoint":true,"statusBadge":"CRITICAL CYBERSECURITY","category":"reasoning","categoryLabel":"Cybersecurity Frontier","contextWindow":"1,050,000 tokens","contextWindowTokens":1050000,"maxOutputTokens":"65,536 tokens","parameters":"Universal Chain-of-Thought","openWeights":false,"license":"Restricted Defense / Preparedness API","pricing":{"input":15,"output":60},"highlight":"Announced Sept 1, 2026; the first LLM to meet OpenAI's 'Critical' cybersecurity threshold under its Preparedness Framework, capable of identifying zero-day exploits.","modalities":["Text","Code"],"benchmarks":{"sweBench":"77.5%","gpqa":"75.1%"},"links":{"announcement":"https://openai.com/index/preparedness-framework-astra/"}},{"id":"gpt-5-6-luna","companyId":"openai","companyName":"OpenAI","name":"GPT-5.6 Luna","version":"5.6-Luna","releaseDate":"2026-07-28","isCompanyFlagship":false,"isLatestCheckpoint":false,"statusBadge":"HIGH THROUGHPUT","category":"code","categoryLabel":"Fast Inference","contextWindow":"256,000 tokens","contextWindowTokens":256000,"maxOutputTokens":"16,384 tokens","parameters":"Distilled High-Throughput","openWeights":false,"license":"OpenAI API","pricing":{"input":0.2,"output":0.8},"highlight":"High-speed developer model designed for streaming code completions, real-time function calling, and high-frequency queries at $0.20/1M tokens.","modalities":["Text","Code"],"benchmarks":{"mmluPro":"79.1%"},"links":{"apiDocs":"https://platform.openai.com/docs"}},{"id":"llama-4-maverick","companyId":"meta","companyName":"Meta AI","name":"Llama 4 Maverick (128E)","version":"4.0","releaseDate":"2026-04-18","isCompanyFlagship":true,"isLatestCheckpoint":false,"statusBadge":"OPEN WEIGHTS SOTA","category":"open-weights","categoryLabel":"128-Expert Open MoE","contextWindow":"1,048,576 tokens","contextWindowTokens":1048576,"maxOutputTokens":"32,768 tokens","parameters":"128-Expert Mixture-of-Experts","openWeights":true,"license":"Meta Community License","pricing":{"input":0.45,"output":1.35},"highlight":"Meta's flagship open-weights foundation model. 128-expert MoE architecture with 1M token context, powering enterprise on-premise deployments and community fine-tuning.","modalities":["Text","Vision","Code"],"benchmarks":{"mmluPro":"83.6%","sweBench":"71.2%"},"links":{"announcement":"https://ai.meta.com/blog/llama-4/","weights":"https://huggingface.co/meta-llama"}},{"id":"meta-muse-voice-transcribe","companyId":"meta","companyName":"Meta AI","name":"Muse Voice Transcribe","version":"1.0","releaseDate":"2026-09-01","isCompanyFlagship":false,"isLatestCheckpoint":true,"statusBadge":"NEW AUDIO CHECKPOINT","category":"audio","categoryLabel":"Streaming Speech","contextWindow":"128,000 tokens","contextWindowTokens":128000,"maxOutputTokens":"16,384 tokens","parameters":"Streaming Audio Foundation","openWeights":true,"license":"Meta Community License","pricing":{"input":0.1,"output":0.3},"highlight":"Announced Sept 1, 2026; streaming speech-to-text foundation model supporting 70+ languages, 20+ voice diarization, and multilingual code-switching.","modalities":["Audio","Text"],"benchmarks":{},"links":{"announcement":"https://ai.meta.com/blog/muse-voice-transcribe/"}},{"id":"llama-4-scout","companyId":"meta","companyName":"Meta AI","name":"Llama 4 Scout (16E)","version":"4.0-Scout","releaseDate":"2026-05-10","isCompanyFlagship":false,"isLatestCheckpoint":false,"statusBadge":"LONG CONTEXT MOE","category":"open-weights","categoryLabel":"1.31M Context Scout","contextWindow":"1,310,720 tokens","contextWindowTokens":1310720,"maxOutputTokens":"32,768 tokens","parameters":"16-Expert MoE","openWeights":true,"license":"Meta Community License","pricing":{"input":0.2,"output":0.6},"highlight":"Extended-context open-weights model capable of ingesting 1.31 million tokens in a single prompt for codebase-wide document analysis.","modalities":["Text","Code"],"benchmarks":{"mmluPro":"78.4%"},"links":{"weights":"https://huggingface.co/meta-llama"}},{"id":"gemini-3-8-flash","companyId":"google","companyName":"Google DeepMind","name":"Gemini 3.8 Flash","version":"3.8","releaseDate":"2026-09-02","isCompanyFlagship":true,"isLatestCheckpoint":true,"statusBadge":"NEWEST FLASH SOTA","category":"flagship","categoryLabel":"Agentic Multimodal","contextWindow":"1,048,576 tokens","contextWindowTokens":1048576,"maxOutputTokens":"65,536 tokens","parameters":"TPU v6e Speculative MoE","openWeights":false,"license":"Google AI Studio / Vertex AI","pricing":{"input":0.75,"output":3.75},"highlight":"Released Sept 2, 2026; Google DeepMind's premier frontier workhorse powering advanced agentic coding loops, recursive self-correction, and 1M token real-time multimodal streaming.","modalities":["Text","Vision","Audio","Video","Code"],"benchmarks":{"mmluPro":"86.8%","sweBench":"76.4%","gpqa":"73.2%"},"links":{"announcement":"https://blog.google/technology/ai/gemini-3-8-flash/","playground":"https://aistudio.google.com"}},{"id":"deepseek-v4-pro-0813","companyId":"deepseek","companyName":"DeepSeek","name":"DeepSeek V4-Pro (0813)","version":"V4-Pro","releaseDate":"2026-08-13","isCompanyFlagship":true,"isLatestCheckpoint":false,"statusBadge":"1.6T HOSTED SOTA","category":"flagship","categoryLabel":"Frontier MoE","contextWindow":"1,048,576 tokens","contextWindowTokens":1048576,"maxOutputTokens":"65,536 tokens","parameters":"1.6 Trillion Total (49B Activated)","openWeights":false,"license":"DeepSeek API / Enterprise","pricing":{"input":1.12,"output":3.35},"highlight":"DeepSeek's primary 1.6T parameter powerhouse with 49B activated per token, configurable thinking budget, and 1M context.","modalities":["Text","Code"],"benchmarks":{"mmluPro":"85.7%","sweBench":"72.4%"},"links":{"announcement":"https://api-docs.deepseek.com/news/news260813","playground":"https://chat.deepseek.com"}},{"id":"deepseek-v4-flash-vision-exp","companyId":"deepseek","companyName":"DeepSeek","name":"DeepSeek V4 Flash Vision Exp","version":"V4-Vision-Exp","releaseDate":"2026-08-30","isCompanyFlagship":false,"isLatestCheckpoint":true,"statusBadge":"OPEN WEIGHTS (MIT)","category":"open-weights","categoryLabel":"Open Vision MoE","contextWindow":"262,144 tokens","contextWindowTokens":262144,"maxOutputTokens":"32,768 tokens","parameters":"305B MoE","openWeights":true,"license":"MIT License","pricing":{"input":0.12,"output":0.36},"highlight":"Open-sourced under MIT license on August 30, 2026. 305B parameter multimodal vision-language model with native document understanding.","modalities":["Text","Vision","Code"],"benchmarks":{"mmluPro":"81.4%"},"links":{"weights":"https://huggingface.co/deepseek-ai"}},{"id":"grok-4-6","companyId":"xai","companyName":"xAI","name":"Grok 4.6","version":"4.6","releaseDate":"2026-08-12","isCompanyFlagship":true,"isLatestCheckpoint":true,"statusBadge":"1.5T SUPERCOMPUTE","category":"flagship","categoryLabel":"Frontier Coding & STEM","contextWindow":"500,000 tokens","contextWindowTokens":500000,"maxOutputTokens":"65,536 tokens","parameters":"1.5 Trillion Parameters","openWeights":false,"license":"Proprietary API / Grok Build","pricing":{"input":2,"output":6},"highlight":"Released August 12, 2026; xAI's smartest model with frontier performance in coding and autonomous agents, integrated natively into Cursor and Grok Build.","modalities":["Text","Vision","Code"],"benchmarks":{"mmluPro":"86.1%","sweBench":"75.8%"},"links":{"announcement":"https://x.ai/blog/grok-4-6","playground":"https://x.com/i/grok"}},{"id":"qwen-3-8-2-4t-a95b","companyId":"qwen","companyName":"Alibaba Cloud (Qwen)","name":"Qwen3.8 2.4T A95B","version":"3.8-2.4T","releaseDate":"2026-08-12","isCompanyFlagship":true,"isLatestCheckpoint":false,"statusBadge":"2.4T OPEN TITAN","category":"open-weights","categoryLabel":"Largest Open MoE","contextWindow":"1,000,000 tokens","contextWindowTokens":1000000,"maxOutputTokens":"32,768 tokens","parameters":"2.4 Trillion Total (95B Activated)","openWeights":true,"license":"Qwen Community License","pricing":{"input":1.8,"output":5.4},"highlight":"The largest open-weight MoE model in existence. 2.4 Trillion parameters with 95B activated per token and 1M context, available on HuggingFace and ModelScope.","modalities":["Text","Code"],"benchmarks":{"mmluPro":"85.2%","sweBench":"74.1%"},"links":{"announcement":"https://qwenlm.github.io/blog/qwen3.8-2.4t/","weights":"https://huggingface.co/Qwen"}},{"id":"qwen-3-8-flash","companyId":"qwen","companyName":"Alibaba Cloud (Qwen)","name":"Qwen3.8 Flash","version":"3.8-Flash","releaseDate":"2026-08-26","isCompanyFlagship":false,"isLatestCheckpoint":true,"statusBadge":"1M MULTIMODAL VALUE","category":"multimodal","categoryLabel":"Fast Multimodal","contextWindow":"1,000,000 tokens","contextWindowTokens":1000000,"maxOutputTokens":"32,768 tokens","parameters":"Next-Gen Qwen MoE","openWeights":false,"license":"Alibaba Cloud Model Studio","pricing":{"input":0.15,"output":0.47},"highlight":"Released August 26, 2026; combines visual document understanding, fast agentic workflows, and 1M context with ultra-cheap $0.15/$0.47 pricing.","modalities":["Text","Vision","Code"],"benchmarks":{"mmluPro":"81.9%"},"links":{"announcement":"https://qwenlm.github.io/blog/qwen3.8/","playground":"https://chat.qwenlm.ai"}},{"id":"cohere-command-a","companyId":"cohere","companyName":"Cohere","name":"Command A (111B)","version":"1.0","releaseDate":"2026-03-20","isCompanyFlagship":true,"isLatestCheckpoint":false,"statusBadge":"ENTERPRISE FLAGSHIP","category":"flagship","categoryLabel":"Enterprise RAG","contextWindow":"256,000 tokens","contextWindowTokens":256000,"maxOutputTokens":"16,384 tokens","parameters":"111B Parameters","openWeights":true,"license":"CC-BY-NC 4.0 / Cohere API","pricing":{"input":1.25,"output":5},"highlight":"Cohere's primary flagship model optimized for business agentic search, long-document question answering, and multi-step tool use.","modalities":["Text","Code"],"benchmarks":{"mmluPro":"77.8%"},"links":{"announcement":"https://cohere.com/blog/command-a","weights":"https://huggingface.co/CohereForAI"}},{"id":"cohere-north-mini-code","companyId":"cohere","companyName":"Cohere","name":"North Mini Code","version":"1.0","releaseDate":"2026-06-17","isCompanyFlagship":false,"isLatestCheckpoint":true,"statusBadge":"FREE AGENTIC CODE","category":"code","categoryLabel":"Autonomous Coding","contextWindow":"256,000 tokens","contextWindowTokens":256000,"maxOutputTokens":"16,384 tokens","parameters":"30B MoE (North Architecture)","openWeights":false,"license":"Cohere Free / Developer API","pricing":{"input":0,"output":0},"highlight":"30B MoE agentic coding model specifically tuned for multi-file workspace inspection and test suites, offered free for developers.","modalities":["Text","Code"],"benchmarks":{"sweBench":"64.1%"},"links":{"announcement":"https://cohere.com/blog/north-mini-code","apiDocs":"https://docs.cohere.com"}},{"id":"mistral-medium-3-5","companyId":"mistral","companyName":"Mistral AI","name":"Mistral Medium 3.5","version":"3.5","releaseDate":"2026-04-30","isCompanyFlagship":true,"isLatestCheckpoint":true,"statusBadge":"EUROPEAN FLAGSHIP","category":"flagship","categoryLabel":"Dense Multimodal","contextWindow":"262,144 tokens","contextWindowTokens":262144,"maxOutputTokens":"16,384 tokens","parameters":"128B Dense Multimodal","openWeights":true,"license":"Mistral Commercial API / Research","pricing":{"input":1.5,"output":7.5},"highlight":"Dense 128B multimodal instruction-following model with native text and image understanding, tuned for European enterprise compliance.","modalities":["Text","Vision","Code"],"benchmarks":{"mmluPro":"79.2%"},"links":{"announcement":"https://mistral.ai/news/mistral-medium-3-5/"}}]}