{ "name": "AI Model Release Timeline: Every Major LLM From GPT-1 to GPT-6 (2018-2026)", "description": "AI model release timeline of 159 LLMs from GPT-1 to GPT-6: release dates, open weights, context windows and launch API prices, each linked to the lab's source.", "url": "https://agitracker.io/data/ai-model-release-timeline/", "license": "https://creativecommons.org/licenses/by/4.0/", "publisher": "agitracker.io", "updated": "2026-10-29", "columns": [ { "key": "model", "label": "Model" }, { "key": "lab", "label": "Lab" }, { "key": "released", "label": "Released", "type": "date", "note": "First public availability (app, API or weights)" }, { "key": "openWeights", "label": "Open weights", "note": "Yes = weights downloadable, under any license" }, { "key": "context", "label": "Context window", "type": "number", "note": "Tokens at launch; max sequence length for early models" }, { "key": "modalities", "label": "Input types", "type": "tags" }, { "key": "inputPrice", "label": "Input price", "type": "number", "note": "USD per 1M tokens, lab's own API, standard list price at launch" }, { "key": "outputPrice", "label": "Output price", "type": "number", "note": "USD per 1M tokens, lab's own API, standard list price at launch" }, { "key": "notes", "label": "Notes" } ], "rows": [ { "model": "GPT-1", "lab": "OpenAI", "released": "2018-06-11", "openWeights": "Yes", "context": 512, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "First GPT model: generative pre-training of a Transformer followed by fine-tuning", "source": { "title": "OpenAI: Improving language understanding with unsupervised learning", "url": "https://openai.com/index/language-unsupervised/" } }, { "model": "BERT", "lab": "Google", "released": "2018-11-02", "openWeights": "Yes", "context": 512, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Bidirectional transformer encoder; Google open-sourced code and pre-trained models", "source": { "title": "Google Research blog: Open Sourcing BERT", "url": "https://research.google/blog/open-sourcing-bert-state-of-the-art-pre-training-for-natural-language-processing/" } }, { "model": "GPT-2", "lab": "OpenAI", "released": "2019-02-14", "openWeights": "Yes", "context": 1024, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Staged release: only a smaller model at first; full 1.5B-parameter weights published 2019-11-05", "source": { "title": "OpenAI: Better language models and their implications", "url": "https://openai.com/index/better-language-models/" } }, { "model": "T5", "lab": "Google", "released": "2019-10-23", "openWeights": "Yes", "context": null, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Text-to-text framework; largest model 11B parameters; released with the C4 dataset", "source": { "title": "arXiv: Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer", "url": "https://arxiv.org/abs/1910.10683" } }, { "model": "GPT-3", "lab": "OpenAI", "released": "2020-06-11", "openWeights": "No", "context": 2048, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "175 billion parameters; first OpenAI model sold through an API (private beta)", "source": { "title": "OpenAI: OpenAI API", "url": "https://openai.com/index/openai-api/" } }, { "model": "Jurassic-1 Jumbo", "lab": "AI21 Labs", "released": "2021-08", "openWeights": "No", "context": null, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "178B parameters; offered through the AI21 Studio open beta API", "source": { "title": "AI21 Labs: Announcing AI21 Studio and Jurassic-1", "url": "https://www.ai21.com/blog/announcing-ai21-studio-and-jurassic-1" } }, { "model": "GPT-NeoX-20B", "lab": "EleutherAI", "released": "2022-02-09", "openWeights": "Yes", "context": 2048, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "20B parameters; weights released under Apache 2.0 after a one-week delay from the Feb 2 announcement.", "source": { "title": "Announcing GPT-NeoX-20B (EleutherAI blog)", "url": "https://blog.eleuther.ai/announcing-20b/" } }, { "model": "OPT-175B", "lab": "Meta", "released": "2022-05-03", "openWeights": "Yes", "context": 2048, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "175B parameters; weights by request under a noncommercial research license (OPT-175B License).", "source": { "title": "Democratizing access to large-scale language models with OPT-175B (Meta AI)", "url": "https://ai.meta.com/blog/democratizing-access-to-large-scale-language-models-with-opt-175b/" } }, { "model": "BLOOM", "lab": "BigScience", "released": "2022-07-11", "openWeights": "Yes", "context": 2048, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "176B-parameter multilingual model built by a 1,000+ researcher collaboration; BigScience RAIL License v1.0.", "source": { "title": "bigscience/bloom model card (Hugging Face)", "url": "https://huggingface.co/bigscience/bloom" } }, { "model": "ChatGPT (GPT-3.5)", "lab": "OpenAI", "released": "2022-11-30", "openWeights": "No", "context": null, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Free research preview of a chat interface trained with RLHF; no API until March 2023", "source": { "title": "OpenAI: Introducing ChatGPT", "url": "https://openai.com/index/chatgpt/" } }, { "model": "LLaMA", "lab": "Meta", "released": "2023-02-24", "openWeights": "Yes", "context": 2048, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "7B to 65B parameters; weights granted case by case under a noncommercial research license.", "source": { "title": "Introducing LLaMA: A foundational, 65-billion-parameter large language model (Meta AI)", "url": "https://ai.meta.com/blog/large-language-model-llama-meta-ai/" } }, { "model": "GPT-3.5 Turbo (ChatGPT API)", "lab": "OpenAI", "released": "2023-03-01", "openWeights": "No", "context": null, "modalities": [ "Text" ], "inputPrice": 2, "outputPrice": 2, "notes": "Same model as ChatGPT; priced at $0.002 per 1K tokens, 10x cheaper than earlier GPT-3.5 models", "source": { "title": "OpenAI: Introducing ChatGPT and Whisper APIs", "url": "https://openai.com/index/introducing-chatgpt-and-whisper-apis/" } }, { "model": "Claude", "lab": "Anthropic", "released": "2023-03-14", "openWeights": "No", "context": null, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "First Claude release, alongside the lighter Claude Instant; offered via chat interface and API", "source": { "title": "Anthropic: Introducing Claude", "url": "https://www.anthropic.com/news/introducing-claude" } }, { "model": "GPT-4", "lab": "OpenAI", "released": "2023-03-14", "openWeights": "No", "context": 8192, "modalities": [ "Text" ], "inputPrice": 30, "outputPrice": 60, "notes": "32,768-token gpt-4-32k variant at $60/$120; image input demonstrated but not in the API at launch", "source": { "title": "OpenAI: GPT-4", "url": "https://openai.com/index/gpt-4-research/" } }, { "model": "PaLM 2", "lab": "Google", "released": "2023-05-10", "openWeights": "No", "context": null, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Announced at Google I/O in four sizes (Gecko, Otter, Bison, Unicorn); powered Bard and 25+ Google products", "source": { "title": "Google: Introducing PaLM 2", "url": "https://blog.google/technology/ai/google-palm-2-ai-large-language-model/" } }, { "model": "Falcon 40B", "lab": "TII (UAE)", "released": "2023-05-25", "openWeights": "Yes", "context": 2048, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "40B parameters trained on 1 trillion tokens; made royalty-free under Apache 2.0 on May 31, 2023.", "source": { "title": "UAE's TII Launches Open-Source Falcon 40B (TII)", "url": "https://www.tii.ae/news/uaes-technology-innovation-institute-launches-open-source-falcon-40b-large-language-model" } }, { "model": "Claude 2", "lab": "Anthropic", "released": "2023-07-11", "openWeights": "No", "context": 100000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "100K-token prompts; launched public beta of claude.ai in the US and UK", "source": { "title": "Anthropic: Claude 2", "url": "https://www.anthropic.com/news/claude-2" } }, { "model": "Llama 2", "lab": "Meta", "released": "2023-07-18", "openWeights": "Yes", "context": 4096, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "7B to 70B; first Llama licensed for commercial use (Llama 2 Community License), launched with Microsoft.", "source": { "title": "Meta and Microsoft Introduce the Next Generation of Llama (Meta AI)", "url": "https://ai.meta.com/blog/llama-2/" } }, { "model": "Qwen-7B", "lab": "Alibaba Qwen", "released": "2023-08-03", "openWeights": "Yes", "context": 2048, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "First open-weight Qwen model, released with Qwen-7B-Chat under the Tongyi Qianwen license; context later extended to 8K in a Sept 2023 update.", "source": { "title": "Qwen GitHub README (news and updates)", "url": "https://github.com/QwenLM/Qwen" } }, { "model": "Code Llama", "lab": "Meta", "released": "2023-08-24", "openWeights": "Yes", "context": 100000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Code-specialized Llama 2; trained on 16K-token sequences, stable up to 100K tokens; 70B version followed Jan 29, 2024.", "source": { "title": "Introducing Code Llama (Meta AI)", "url": "https://ai.meta.com/blog/code-llama-large-language-model-coding/" } }, { "model": "Falcon 180B", "lab": "TII (UAE)", "released": "2023-09-06", "openWeights": "Yes", "context": null, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "180B parameters trained on 3.5 trillion tokens; Falcon 180B TII License based on Apache 2.0.", "source": { "title": "TII Introduces World's Most Powerful Open LLM: Falcon 180B (TII)", "url": "https://www.tii.ae/news/technology-innovation-institute-introduces-worlds-most-powerful-open-llm-falcon-180b" } }, { "model": "Mistral 7B", "lab": "Mistral AI", "released": "2023-09-27", "openWeights": "Yes", "context": null, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Mistral's first model; Apache 2.0, uses sliding window attention.", "source": { "title": "Mistral 7B (Mistral AI)", "url": "https://mistral.ai/news/announcing-mistral-7b" } }, { "model": "Yi-34B", "lab": "01.AI", "released": "2023-11-02", "openWeights": "Yes", "context": 4096, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "34B dense model; a 200K-context variant (Yi-34B-200K) followed on 2023-11-05. Model card now lists Apache 2.0.", "source": { "title": "Yi-34B model card (01.AI, Hugging Face)", "url": "https://huggingface.co/01-ai/Yi-34B" } }, { "model": "GPT-4 Turbo", "lab": "OpenAI", "released": "2023-11-06", "openWeights": "No", "context": 128000, "modalities": [ "Text", "Image" ], "inputPrice": 10, "outputPrice": 30, "notes": "First OpenAI model with a 128K context window; launched as a preview at DevDay", "source": { "title": "OpenAI: New models and developer products announced at DevDay", "url": "https://openai.com/index/new-models-and-developer-products-announced-at-devday/" } }, { "model": "Claude 2.1", "lab": "Anthropic", "released": "2023-11-21", "openWeights": "No", "context": 200000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Doubled context to 200K tokens; added system prompts and tool use in beta", "source": { "title": "Anthropic: Introducing Claude 2.1", "url": "https://www.anthropic.com/news/claude-2-1" } }, { "model": "Amazon Titan Text Express", "lab": "Amazon", "released": "2023-11-29", "openWeights": "No", "context": 8000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Generally available on Amazon Bedrock with Titan Text Lite (4,000 tokens)", "source": { "title": "AWS What's New: Amazon Titan Text models Express and Lite now generally available", "url": "https://aws.amazon.com/about-aws/whats-new/2023/11/amazon-titan-models-express-lite-bedrock/" } }, { "model": "Gemini 1.0 Pro", "lab": "Google", "released": "2023-12-06", "openWeights": "No", "context": 32000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "First Gemini model; Bard switched to Gemini Pro on launch day; API access from December 13 with a 32K context window; a separate Pro Vision endpoint accepted images", "source": { "title": "Google: Introducing Gemini", "url": "https://blog.google/technology/ai/google-gemini-ai/" } }, { "model": "Mixtral 8x7B", "lab": "Mistral AI", "released": "2023-12-11", "openWeights": "Yes", "context": 32000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Sparse mixture-of-experts model with open weights under Apache 2.0.", "source": { "title": "Mixtral of experts (Mistral AI)", "url": "https://mistral.ai/news/mixtral-of-experts" } }, { "model": "Phi-2", "lab": "Microsoft", "released": "2023-12-12", "openWeights": "Yes", "context": 2048, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "2.7B-parameter small language model; MIT license", "source": { "title": "Microsoft Research: Phi-2: The surprising power of small language models", "url": "https://www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/" } }, { "model": "OLMo 7B", "lab": "Ai2", "released": "2024-02-01", "openWeights": "Yes", "context": 2048, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Released weights plus full training data, code, logs and checkpoints under Apache 2.0.", "source": { "title": "OLMo: Open Language Model (Ai2)", "url": "https://allenai.org/blog/olmo-open-language-model-87ccfc95f580" } }, { "model": "Qwen1.5", "lab": "Alibaba Qwen", "released": "2024-02-04", "openWeights": "Yes", "context": 32768, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Family from 0.5B to 72B (later 110B); all sizes support a 32,768-token context.", "source": { "title": "Introducing Qwen1.5 (Qwen blog)", "url": "https://qwenlm.github.io/blog/qwen1.5/" } }, { "model": "Gemma", "lab": "Google", "released": "2024-02-21", "openWeights": "Yes", "context": 8192, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Google's first open-weight Gemma models (2B and 7B)", "source": { "title": "Google: Gemma, introducing new state-of-the-art open models", "url": "https://blog.google/technology/developers/gemma-open-models/" } }, { "model": "Mistral Large", "lab": "Mistral AI", "released": "2024-02-26", "openWeights": "No", "context": 32000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Closed flagship model offered through Mistral's API.", "source": { "title": "Au Large (Mistral AI)", "url": "https://mistral.ai/news/mistral-large" } }, { "model": "Claude 3 Opus", "lab": "Anthropic", "released": "2024-03-04", "openWeights": "No", "context": 200000, "modalities": [ "Text", "Image" ], "inputPrice": 15, "outputPrice": 75, "notes": "Claude 3 family launch (Opus, Sonnet); first Claude models with image input", "source": { "title": "Anthropic: Introducing the next generation of Claude", "url": "https://www.anthropic.com/news/claude-3-family" } }, { "model": "Command R", "lab": "Cohere", "released": "2024-03-11", "openWeights": "Yes", "context": 128000, "modalities": [ "Text" ], "inputPrice": 0.5, "outputPrice": 1.5, "notes": "RAG- and tool-use-focused model; weights published under a non-commercial CC-BY-NC license", "source": { "title": "Cohere docs: Command R", "url": "https://docs.cohere.com/docs/command-r" } }, { "model": "Claude 3 Haiku", "lab": "Anthropic", "released": "2024-03-13", "openWeights": "No", "context": 200000, "modalities": [ "Text", "Image" ], "inputPrice": 0.25, "outputPrice": 1.25, "notes": "Smallest Claude 3 model", "source": { "title": "Anthropic: Claude 3 Haiku", "url": "https://www.anthropic.com/news/claude-3-haiku" } }, { "model": "Grok-1 (open weights)", "lab": "xAI", "released": "2024-03-17", "openWeights": "Yes", "context": null, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "314B-parameter mixture-of-experts base model; weights and architecture released under Apache 2.0.", "source": { "title": "Open Release of Grok-1 (xAI)", "url": "https://x.ai/news/grok-os" } }, { "model": "Jamba", "lab": "AI21 Labs", "released": "2024-03-28", "openWeights": "Yes", "context": 256000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Hybrid Mamba-Transformer mixture-of-experts model (52B total, 12B active); Apache 2.0 weights", "source": { "title": "AI21 Labs: Introducing Jamba", "url": "https://www.ai21.com/blog/announcing-jamba" } }, { "model": "Command R+", "lab": "Cohere", "released": "2024-04-04", "openWeights": "Yes", "context": 128000, "modalities": [ "Text" ], "inputPrice": 3, "outputPrice": 15, "notes": "Larger RAG model, first available on Microsoft Azure; CC-BY-NC weights", "source": { "title": "Cohere: Introducing Command R+", "url": "https://cohere.com/blog/command-r-plus-microsoft-azure" } }, { "model": "Gemini 1.5 Pro", "lab": "Google", "released": "2024-04-09", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image", "Audio", "Video" ], "inputPrice": null, "outputPrice": null, "notes": "First public API model with a 1M-token context window; announced February 15, 2024 for limited testing", "source": { "title": "Google Developers Blog: Gemini 1.5 Pro now available in 180+ countries", "url": "https://developers.googleblog.com/en/gemini-15-pro-now-available-in-180-countries-with-native-audio-understanding-system-instructions-json-mode-and-more/" } }, { "model": "Mixtral 8x22B", "lab": "Mistral AI", "released": "2024-04-17", "openWeights": "Yes", "context": 64000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Sparse mixture-of-experts; Apache 2.0.", "source": { "title": "Cheaper, Better, Faster, Stronger (Mistral AI)", "url": "https://mistral.ai/news/mixtral-8x22b" } }, { "model": "Llama 3", "lab": "Meta", "released": "2024-04-18", "openWeights": "Yes", "context": 8192, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "8B and 70B; trained on 8,192-token sequences with a new 128K-vocabulary tokenizer.", "source": { "title": "Introducing Meta Llama 3 (Meta AI)", "url": "https://ai.meta.com/blog/meta-llama-3/" } }, { "model": "Phi-3-mini", "lab": "Microsoft", "released": "2024-04-23", "openWeights": "Yes", "context": 128000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "3.8B parameters; 4K and 128K context variants; Microsoft says first model in its class with a 128K window", "source": { "title": "Microsoft Azure Blog: Introducing Phi-3", "url": "https://azure.microsoft.com/en-us/blog/introducing-phi-3-redefining-whats-possible-with-slms/" } }, { "model": "DeepSeek-V2", "lab": "DeepSeek", "released": "2024-05-06", "openWeights": "Yes", "context": 128000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Mixture-of-experts model with 236B total and 21B active parameters; introduced Multi-head Latent Attention. DeepSeek Model License (commercial use allowed).", "source": { "title": "DeepSeek-V2 GitHub README", "url": "https://github.com/deepseek-ai/DeepSeek-V2" } }, { "model": "GPT-4o", "lab": "OpenAI", "released": "2024-05-13", "openWeights": "No", "context": 128000, "modalities": [ "Text", "Image" ], "inputPrice": 5, "outputPrice": 15, "notes": "'Omni' model; API launched with text and vision, audio added later", "source": { "title": "OpenAI API changelog (May 13, 2024)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "Codestral", "lab": "Mistral AI", "released": "2024-05-29", "openWeights": "Yes", "context": 32000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Mistral's first code model, 22B; Mistral AI Non-Production License.", "source": { "title": "Codestral (Mistral AI)", "url": "https://mistral.ai/news/codestral" } }, { "model": "Qwen2", "lab": "Alibaba Qwen", "released": "2024-06-07", "openWeights": "Yes", "context": 131072, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Sizes 0.5B to 72B; Qwen2-72B kept the Qianwen license while smaller models moved to Apache 2.0.", "source": { "title": "Hello Qwen2 (Qwen blog)", "url": "https://qwenlm.github.io/blog/qwen2/" } }, { "model": "Nemotron-4 340B", "lab": "NVIDIA", "released": "2024-06-14", "openWeights": "Yes", "context": null, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Family of open models aimed at synthetic data generation; NVIDIA Open Model License.", "source": { "title": "NVIDIA Releases Open Synthetic Data Generation Pipeline (NVIDIA blog)", "url": "https://blogs.nvidia.com/blog/nemotron-4-synthetic-data-generation-llm-training/" } }, { "model": "Claude 3.5 Sonnet", "lab": "Anthropic", "released": "2024-06-20", "openWeights": "No", "context": 200000, "modalities": [ "Text", "Image" ], "inputPrice": 3, "outputPrice": 15, "notes": "Introduced Artifacts in claude.ai; upgraded version in October 2024 added computer use in beta", "source": { "title": "Anthropic: Introducing Claude 3.5 Sonnet", "url": "https://www.anthropic.com/news/claude-3-5-sonnet" } }, { "model": "Gemma 2", "lab": "Google", "released": "2024-06-27", "openWeights": "Yes", "context": 8192, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "9B and 27B open models", "source": { "title": "Google: Gemma 2 is now available to researchers and developers", "url": "https://blog.google/technology/developers/google-gemma-2/" } }, { "model": "GPT-4o mini", "lab": "OpenAI", "released": "2024-07-18", "openWeights": "No", "context": 128000, "modalities": [ "Text", "Image" ], "inputPrice": 0.15, "outputPrice": 0.6, "notes": "Small, low-cost model that replaced GPT-3.5 Turbo as the cheap default", "source": { "title": "OpenAI API changelog (Jul 18, 2024)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "Llama 3.1 405B", "lab": "Meta", "released": "2024-07-23", "openWeights": "Yes", "context": 128000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Meta called the 405B model the first frontier-level open source AI model; 128K context, eight languages.", "source": { "title": "Introducing Llama 3.1: Our most capable models to date (Meta AI)", "url": "https://ai.meta.com/blog/meta-llama-3-1/" } }, { "model": "Mistral Large 2", "lab": "Mistral AI", "released": "2024-07-24", "openWeights": "Yes", "context": 128000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "123B parameters; weights under the Mistral Research License (non-commercial).", "source": { "title": "Large Enough (Mistral AI)", "url": "https://mistral.ai/news/mistral-large-2407" } }, { "model": "Grok-2", "lab": "xAI", "released": "2024-08-13", "openWeights": "No", "context": null, "modalities": [ "Text", "Image" ], "inputPrice": null, "outputPrice": null, "notes": "Released in beta on X with Grok-2 mini; text and vision understanding.", "source": { "title": "Grok-2 Beta Release (xAI)", "url": "https://x.ai/news/grok-2" } }, { "model": "o1-preview", "lab": "OpenAI", "released": "2024-09-12", "openWeights": "No", "context": 128000, "modalities": [ "Text" ], "inputPrice": 15, "outputPrice": 60, "notes": "First OpenAI reasoning model trained with reinforcement learning to think before answering", "source": { "title": "OpenAI API changelog (Sep 12, 2024)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "Qwen2.5", "lab": "Alibaba Qwen", "released": "2024-09-19", "openWeights": "Yes", "context": 131072, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Sizes 0.5B to 72B; most under Apache 2.0, with the 3B and 72B models under Qwen licenses.", "source": { "title": "Qwen2.5: A Party of Foundation Models! (Qwen blog)", "url": "https://qwenlm.github.io/blog/qwen2.5/" } }, { "model": "Llama 3.2 Vision (11B, 90B)", "lab": "Meta", "released": "2024-09-25", "openWeights": "Yes", "context": 128000, "modalities": [ "Text", "Image" ], "inputPrice": null, "outputPrice": null, "notes": "First Llama models with image input; released alongside 1B and 3B text models for devices.", "source": { "title": "Llama 3.2: edge AI and vision with open, customizable models (Meta AI)", "url": "https://ai.meta.com/blog/llama-3-2-connect-2024-vision-edge-mobile-devices/" } }, { "model": "Hunyuan-Large", "lab": "Tencent", "released": "2024-11-05", "openWeights": "Yes", "context": 128000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "MoE with 389B total and 52B active parameters; Tencent described it as the largest open Transformer MoE at release. Instruct model 128K, pretrain 256K.", "source": { "title": "Tencent-Hunyuan-Large GitHub README", "url": "https://github.com/Tencent/Tencent-Hunyuan-Large" } }, { "model": "Amazon Nova Pro", "lab": "Amazon", "released": "2024-12-03", "openWeights": "No", "context": 300000, "modalities": [ "Text", "Image", "Video" ], "inputPrice": null, "outputPrice": null, "notes": "Launched with Nova Micro (text-only, 128K) and Nova Lite (300K) on Amazon Bedrock", "source": { "title": "AWS News Blog: Introducing Amazon Nova foundation models", "url": "https://aws.amazon.com/blogs/aws/introducing-amazon-nova-frontier-intelligence-and-industry-leading-price-performance/" } }, { "model": "Llama 3.3 70B", "lab": "Meta", "released": "2024-12-06", "openWeights": "Yes", "context": 128000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Text-only 70B model; Llama 3.3 Community License.", "source": { "title": "meta-llama/Llama-3.3-70B-Instruct model card (Hugging Face)", "url": "https://huggingface.co/meta-llama/Llama-3.3-70B-Instruct" } }, { "model": "Gemini 2.0 Flash", "lab": "Google", "released": "2024-12-11", "openWeights": "No", "context": 1048576, "modalities": [ "Text", "Image", "Audio", "Video" ], "inputPrice": null, "outputPrice": null, "notes": "Experimental release December 11, 2024; generally available February 5, 2025", "source": { "title": "Google: Introducing Gemini 2.0", "url": "https://blog.google/technology/google-deepmind/google-gemini-ai-update-december-2024/" } }, { "model": "Phi-4", "lab": "Microsoft", "released": "2024-12-12", "openWeights": "Yes", "context": 16000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "14B-parameter small model focused on math reasoning; MIT license", "source": { "title": "Microsoft Foundry Blog: Introducing Phi-4", "url": "https://techcommunity.microsoft.com/blog/azure-ai-foundry-blog/introducing-phi-4-microsoft%E2%80%99s-newest-small-language-model-specializing-in-comple/4357090" } }, { "model": "o1", "lab": "OpenAI", "released": "2024-12-17", "openWeights": "No", "context": 200000, "modalities": [ "Text", "Image" ], "inputPrice": 15, "outputPrice": 60, "notes": "Full o1 in the API with reasoning_effort control and image input", "source": { "title": "OpenAI API changelog (Dec 17, 2024)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "DeepSeek-V3", "lab": "DeepSeek", "released": "2024-12-26", "openWeights": "Yes", "context": 128000, "modalities": [ "Text" ], "inputPrice": 0.27, "outputPrice": 1.1, "notes": "MoE with 671B total and 37B active parameters, trained on 14.8T tokens. Listed price applied from 2025-02-08; V2 prices applied until then.", "source": { "title": "Introducing DeepSeek-V3 (DeepSeek API news)", "url": "https://api-docs.deepseek.com/news/news1226" } }, { "model": "MiniMax-01 (MiniMax-Text-01)", "lab": "MiniMax", "released": "2025-01-15", "openWeights": "Yes", "context": 1000000, "modalities": [ "Text" ], "inputPrice": 0.2, "outputPrice": 1.1, "notes": "MoE with 456B total and 45.9B active parameters using lightning (linear) attention; MiniMax says it handles up to 4M tokens at inference.", "source": { "title": "MiniMax-01 is Now Open-Source (MiniMax news)", "url": "https://www.minimax.io/news/minimax-01-series-2" } }, { "model": "DeepSeek-R1", "lab": "DeepSeek", "released": "2025-01-20", "openWeights": "Yes", "context": 128000, "modalities": [ "Text" ], "inputPrice": 0.55, "outputPrice": 2.19, "notes": "Open-weight reasoning model trained with large-scale RL, released under MIT with six distilled smaller models; DeepSeek reported o1-level math, code and reasoning.", "source": { "title": "DeepSeek-R1 Release (DeepSeek API news, archived copy)", "url": "https://web.archive.org/web/20250202014805/https://api-docs.deepseek.com/news/news250120" } }, { "model": "Mistral Small 3", "lab": "Mistral AI", "released": "2025-01-30", "openWeights": "Yes", "context": 32000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "24B latency-optimized model; Apache 2.0.", "source": { "title": "Mistral Small 3 (Mistral AI)", "url": "https://mistral.ai/news/mistral-small-3" } }, { "model": "o3-mini", "lab": "OpenAI", "released": "2025-01-31", "openWeights": "No", "context": 200000, "modalities": [ "Text" ], "inputPrice": 1.1, "outputPrice": 4.4, "notes": "Small reasoning model for science, math and coding", "source": { "title": "OpenAI API changelog (Jan 31, 2025)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "Grok 3", "lab": "xAI", "released": "2025-02-19", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": null, "outputPrice": null, "notes": "xAI says context is 8x its previous models; added Think reasoning mode and DeepSearch.", "source": { "title": "Grok 3 Beta: The Age of Reasoning Agents (xAI)", "url": "https://x.ai/news/grok-3" } }, { "model": "Claude 3.7 Sonnet", "lab": "Anthropic", "released": "2025-02-24", "openWeights": "No", "context": null, "modalities": [ "Text", "Image" ], "inputPrice": 3, "outputPrice": 15, "notes": "Anthropic calls it the first hybrid reasoning model on the market (instant or extended thinking); launched with Claude Code preview", "source": { "title": "Anthropic: Claude 3.7 Sonnet and Claude Code", "url": "https://www.anthropic.com/news/claude-3-7-sonnet" } }, { "model": "GPT-4.5", "lab": "OpenAI", "released": "2025-02-27", "openWeights": "No", "context": 128000, "modalities": [ "Text", "Image" ], "inputPrice": 75, "outputPrice": 150, "notes": "Research preview; OpenAI's largest chat model at the time and its most expensive per token; deprecated in April 2025", "source": { "title": "OpenAI API changelog (Feb 27, 2025)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "QwQ-32B", "lab": "Alibaba Qwen", "released": "2025-03-06", "openWeights": "Yes", "context": 131072, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "32B dense reasoning model trained with reinforcement learning, Apache 2.0.", "source": { "title": "QwQ-32B: Embracing the Power of Reinforcement Learning (Qwen blog)", "url": "https://qwenlm.github.io/blog/qwq-32b/" } }, { "model": "Gemma 3", "lab": "Google", "released": "2025-03-12", "openWeights": "Yes", "context": 128000, "modalities": [ "Text", "Image" ], "inputPrice": null, "outputPrice": null, "notes": "Open models from 1B to 27B with text and image input", "source": { "title": "Google: Introducing Gemma 3", "url": "https://blog.google/technology/developers/gemma-3/" } }, { "model": "Command A", "lab": "Cohere", "released": "2025-03-13", "openWeights": "Yes", "context": 256000, "modalities": [ "Text" ], "inputPrice": 2.5, "outputPrice": 10, "notes": "111B parameters, runs on two GPUs; CC-BY-NC weights", "source": { "title": "Cohere docs: Command A", "url": "https://docs.cohere.com/docs/command-a" } }, { "model": "Gemini 2.5 Pro", "lab": "Google", "released": "2025-03-25", "openWeights": "No", "context": 1048576, "modalities": [ "Text", "Image", "Audio", "Video", "PDF" ], "inputPrice": 1.25, "outputPrice": 10, "notes": "Thinking model released as experimental; paid preview pricing from April 4, 2025 ($1.25/$10 for prompts up to 200K tokens)", "source": { "title": "Google: Gemini 2.5, our most intelligent AI model", "url": "https://blog.google/technology/google-deepmind/gemini-model-thinking-updates-march-2025/" } }, { "model": "Llama 4 Maverick", "lab": "Meta", "released": "2025-04-05", "openWeights": "Yes", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": null, "outputPrice": null, "notes": "17B active, 400B total parameters, 128 experts; natively multimodal.", "source": { "title": "The Llama 4 herd (Meta AI)", "url": "https://ai.meta.com/blog/llama-4-multimodal-intelligence/" } }, { "model": "Llama 4 Scout", "lab": "Meta", "released": "2025-04-05", "openWeights": "Yes", "context": 10000000, "modalities": [ "Text", "Image" ], "inputPrice": null, "outputPrice": null, "notes": "First Llama mixture-of-experts models (17B active, 109B total); Meta called its 10M-token context industry-leading.", "source": { "title": "The Llama 4 herd (Meta AI)", "url": "https://ai.meta.com/blog/llama-4-multimodal-intelligence/" } }, { "model": "GPT-4.1", "lab": "OpenAI", "released": "2025-04-14", "openWeights": "No", "context": 1047576, "modalities": [ "Text", "Image" ], "inputPrice": 2, "outputPrice": 8, "notes": "First OpenAI models with a 1M-token context window (with GPT-4.1 mini and nano)", "source": { "title": "OpenAI API changelog (Apr 14, 2025)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "o3", "lab": "OpenAI", "released": "2025-04-16", "openWeights": "No", "context": 200000, "modalities": [ "Text", "Image" ], "inputPrice": 10, "outputPrice": 40, "notes": "Released with o4-mini; price cut 80% to $2/$8 on 2025-06-10", "source": { "title": "OpenAI API changelog (Apr 16, 2025)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "Qwen3 (Qwen3-235B-A22B)", "lab": "Alibaba Qwen", "released": "2025-04-29", "openWeights": "Yes", "context": 128000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Hybrid thinking/non-thinking family; flagship MoE has 235B total and 22B active parameters; Apache 2.0.", "source": { "title": "Qwen3: Think Deeper, Act Faster (Qwen blog)", "url": "https://qwenlm.github.io/blog/qwen3/" } }, { "model": "Amazon Nova Premier", "lab": "Amazon", "released": "2025-04-30", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image", "Video" ], "inputPrice": null, "outputPrice": null, "notes": "Amazon's most capable Nova 1 model and a teacher for model distillation", "source": { "title": "AWS News Blog: Amazon Nova Premier", "url": "https://aws.amazon.com/blogs/aws/amazon-nova-premier-our-most-capable-model-for-complex-tasks-and-teacher-for-model-distillation/" } }, { "model": "Mistral Medium 3", "lab": "Mistral AI", "released": "2025-05-07", "openWeights": "No", "context": null, "modalities": [ "Text", "Image" ], "inputPrice": 0.4, "outputPrice": 2, "notes": "Closed model Mistral pitched at 90% of Claude Sonnet 3.7's benchmark scores for lower cost.", "source": { "title": "Medium is the new large (Mistral AI)", "url": "https://mistral.ai/news/mistral-medium-3" } }, { "model": "Claude Opus 4", "lab": "Anthropic", "released": "2025-05-22", "openWeights": "No", "context": null, "modalities": [ "Text", "Image" ], "inputPrice": 15, "outputPrice": 75, "notes": "Claude 4 launch with Sonnet 4 ($3/$15); Claude Code became generally available", "source": { "title": "Anthropic: Introducing Claude 4", "url": "https://www.anthropic.com/news/claude-4" } }, { "model": "Claude Sonnet 4", "lab": "Anthropic", "released": "2025-05-22", "openWeights": "No", "context": null, "modalities": [ "Text", "Image" ], "inputPrice": 3, "outputPrice": 15, "notes": "Released alongside Opus 4; available to free claude.ai users", "source": { "title": "Anthropic: Introducing Claude 4", "url": "https://www.anthropic.com/news/claude-4" } }, { "model": "ERNIE 4.5 (ERNIE-4.5-300B-A47B)", "lab": "Baidu", "released": "2025-06", "openWeights": "Yes", "context": 131072, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Baidu's first open-weight ERNIE family (ten variants up to 424B total parameters), Apache 2.0; rollout began June 29-30, 2025.", "source": { "title": "China's Baidu open-sources its Ernie AI model (SiliconANGLE)", "url": "https://siliconangle.com/2025/06/29/chinas-baidu-declares-war-openai-others-open-sourcing-ernie-ai-model/" } }, { "model": "Magistral Small / Medium", "lab": "Mistral AI", "released": "2025-06-10", "openWeights": "Yes", "context": 128000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Mistral's first reasoning models; Small (24B) open under Apache 2.0, Medium closed. Context figure is for Small.", "source": { "title": "Magistral (Mistral AI)", "url": "https://mistral.ai/news/magistral" } }, { "model": "MiniMax-M1", "lab": "MiniMax", "released": "2025-06-16", "openWeights": "Yes", "context": 1000000, "modalities": [ "Text" ], "inputPrice": 0.4, "outputPrice": 2.2, "notes": "Hybrid-attention reasoning model (456B total, 45.9B active) with a 1M-token context; price shown is for inputs up to 200K tokens. Apache 2.0.", "source": { "title": "MiniMax-M1 announcement (MiniMax news)", "url": "https://www.minimax.io/news/minimaxm1" } }, { "model": "Grok 4", "lab": "xAI", "released": "2025-07-09", "openWeights": "No", "context": 256000, "modalities": [ "Text", "Image" ], "inputPrice": null, "outputPrice": null, "notes": "Native tool use and real-time search; API context 256,000 tokens.", "source": { "title": "Grok 4 (xAI)", "url": "https://x.ai/news/grok-4" } }, { "model": "Kimi K2", "lab": "Moonshot AI", "released": "2025-07-11", "openWeights": "Yes", "context": 128000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "MoE with 1T total and 32B active parameters, trained with the Muon optimizer; modified MIT license.", "source": { "title": "Kimi K2: Open Agentic Intelligence (Moonshot AI)", "url": "https://www.kimi.com/en/blog/kimi-k2" } }, { "model": "GLM-4.5", "lab": "Z.ai (Zhipu)", "released": "2025-07-28", "openWeights": "Yes", "context": 128000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "MoE with 355B total and 32B active parameters (Air: 106B/12B), MIT license.", "source": { "title": "GLM-4.5: Reasoning, Coding, and Agentic Abilities (Z.ai blog)", "url": "https://z.ai/blog/glm-4.5" } }, { "model": "Grok 2 (open weights)", "lab": "xAI", "released": "2025-08", "openWeights": "Yes", "context": null, "modalities": [ "Text", "Image" ], "inputPrice": null, "outputPrice": null, "notes": "Weights of the 2024 Grok 2 model published under the Grok 2 Community License Agreement.", "source": { "title": "xai-org/grok-2 (Hugging Face)", "url": "https://huggingface.co/xai-org/grok-2" } }, { "model": "gpt-oss-120b", "lab": "OpenAI", "released": "2025-08-05", "openWeights": "Yes", "context": 131072, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "OpenAI's first open-weight language models since GPT-2 (with gpt-oss-20b); Apache 2.0; 117B total, 5.1B active parameters", "source": { "title": "OpenAI: Introducing gpt-oss", "url": "https://openai.com/index/introducing-gpt-oss/" } }, { "model": "GPT-5", "lab": "OpenAI", "released": "2025-08-07", "openWeights": "No", "context": 400000, "modalities": [ "Text", "Image" ], "inputPrice": 1.25, "outputPrice": 10, "notes": "Released in three sizes (gpt-5, mini, nano); 272K input plus 128K output tokens", "source": { "title": "OpenAI: Introducing GPT-5 for developers", "url": "https://openai.com/index/introducing-gpt-5-for-developers/" } }, { "model": "Seed-OSS-36B", "lab": "ByteDance Seed", "released": "2025-08-20", "openWeights": "Yes", "context": 512000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "36B dense model trained on 12T tokens with native 512K context, Apache 2.0.", "source": { "title": "Seed-OSS-36B-Instruct model card (Hugging Face)", "url": "https://huggingface.co/ByteDance-Seed/Seed-OSS-36B-Instruct" } }, { "model": "DeepSeek-V3.1", "lab": "DeepSeek", "released": "2025-08-21", "openWeights": "Yes", "context": 128000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Single model with switchable thinking and non-thinking modes; MIT license.", "source": { "title": "DeepSeek-V3.1 Release (DeepSeek API news)", "url": "https://api-docs.deepseek.com/news/news250821" } }, { "model": "MAI-1-preview", "lab": "Microsoft", "released": "2025-08-28", "openWeights": "No", "context": null, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Microsoft AI's first foundation model trained end to end in-house; public testing on LMArena", "source": { "title": "Microsoft AI: Two in-house models in support of our mission", "url": "https://microsoft.ai/news/two-new-in-house-models/" } }, { "model": "Qwen3-Max-Preview", "lab": "Alibaba Qwen", "released": "2025-09-05", "openWeights": "No", "context": 262144, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Alibaba's first model above 1 trillion parameters; API and Qwen Chat only, no open weights.", "source": { "title": "Qwen3-Max arrives in preview with 1 trillion parameters (VentureBeat)", "url": "https://venturebeat.com/ai/qwen3-max-arrives-in-preview-with-1-trillion-parameters-blazing-fast" } }, { "model": "Claude Sonnet 4.5", "lab": "Anthropic", "released": "2025-09-29", "openWeights": "No", "context": 200000, "modalities": [ "Text", "Image" ], "inputPrice": 3, "outputPrice": 15, "notes": "Sonnet upgrade focused on coding and agents", "source": { "title": "Anthropic: Introducing Claude Sonnet 4.5", "url": "https://www.anthropic.com/news/claude-sonnet-4-5" } }, { "model": "Claude Haiku 4.5", "lab": "Anthropic", "released": "2025-10-15", "openWeights": "No", "context": 200000, "modalities": [ "Text", "Image" ], "inputPrice": 1, "outputPrice": 5, "notes": "Small model priced at $1/$5", "source": { "title": "Anthropic: Introducing Claude Haiku 4.5", "url": "https://www.anthropic.com/news/claude-haiku-4-5" } }, { "model": "MiniMax-M2", "lab": "MiniMax", "released": "2025-10-27", "openWeights": "Yes", "context": null, "modalities": [ "Text" ], "inputPrice": 0.3, "outputPrice": 1.2, "notes": "MoE with 230B total and 10B active parameters aimed at coding and agents; modified MIT license.", "source": { "title": "MiniMax M2 & Agent: Ingenious in Simplicity (MiniMax news)", "url": "https://www.minimax.io/news/minimax-m2" } }, { "model": "Kimi K2 Thinking", "lab": "Moonshot AI", "released": "2025-11-06", "openWeights": "Yes", "context": 256000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Thinking-agent version of K2 (1T total, 32B active) that reasons while calling tools; modified MIT license.", "source": { "title": "Introducing Kimi K2 Thinking (Moonshot AI)", "url": "https://www.kimi.com/en/blog/kimi-k2-thinking" } }, { "model": "GPT-5.1", "lab": "OpenAI", "released": "2025-11-13", "openWeights": "No", "context": 400000, "modalities": [ "Text", "Image" ], "inputPrice": 1.25, "outputPrice": 10, "notes": "Added a 'none' reasoning setting as the default for faster replies", "source": { "title": "OpenAI API changelog (Nov 13, 2025)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "Gemini 3 Pro", "lab": "Google", "released": "2025-11-18", "openWeights": "No", "context": 1048576, "modalities": [ "Text", "Image", "Audio", "Video", "PDF" ], "inputPrice": 2, "outputPrice": 12, "notes": "First Gemini 3 model, released in preview; price applies to prompts up to 200K tokens", "source": { "title": "Google: Gemini 3 for developers", "url": "https://blog.google/technology/developers/gemini-3-developers/" } }, { "model": "Claude Opus 4.5", "lab": "Anthropic", "released": "2025-11-24", "openWeights": "No", "context": 200000, "modalities": [ "Text", "Image" ], "inputPrice": 5, "outputPrice": 25, "notes": "Cut Opus pricing from $15/$75 to $5/$25", "source": { "title": "Anthropic: Introducing Claude Opus 4.5", "url": "https://www.anthropic.com/news/claude-opus-4-5" } }, { "model": "DeepSeek-V3.2", "lab": "DeepSeek", "released": "2025-12-01", "openWeights": "Yes", "context": 128000, "modalities": [ "Text" ], "inputPrice": 0.28, "outputPrice": 0.42, "notes": "Successor to V3.2-Exp (2025-09-29), which introduced DeepSeek Sparse Attention; first DeepSeek model to integrate thinking into tool use. MIT license.", "source": { "title": "DeepSeek-V3.2 Release (DeepSeek API news)", "url": "https://api-docs.deepseek.com/news/news251201" } }, { "model": "Amazon Nova 2 Lite", "lab": "Amazon", "released": "2025-12-02", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image", "Video", "PDF" ], "inputPrice": null, "outputPrice": null, "notes": "Reasoning model with three thinking levels; Nova 2 Pro released the same day in preview", "source": { "title": "AWS News Blog: Introducing Amazon Nova 2 Lite", "url": "https://aws.amazon.com/blogs/aws/introducing-amazon-nova-2-lite-a-fast-cost-effective-reasoning-model/" } }, { "model": "Mistral Large 3", "lab": "Mistral AI", "released": "2025-12-02", "openWeights": "Yes", "context": 256000, "modalities": [ "Text", "Image" ], "inputPrice": 0.5, "outputPrice": 1.5, "notes": "675B total / 41B active mixture-of-experts; Apache 2.0 with image understanding.", "source": { "title": "Introducing Mistral 3 (Mistral AI)", "url": "https://mistral.ai/news/mistral-3" } }, { "model": "GPT-5.2", "lab": "OpenAI", "released": "2025-12-11", "openWeights": "No", "context": 400000, "modalities": [ "Text", "Image" ], "inputPrice": 1.75, "outputPrice": 14, "notes": "Added an xhigh reasoning effort level and compaction", "source": { "title": "OpenAI API changelog (Dec 11, 2025)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "Jamba2", "lab": "AI21 Labs", "released": "2026-01-08", "openWeights": "Yes", "context": 256000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Jamba2 3B and Jamba2 Mini (52B total, 12B active); Apache 2.0 weights", "source": { "title": "AI21 Labs: Introducing Jamba2", "url": "https://www.ai21.com/blog/introducing-jamba2/" } }, { "model": "Kimi K2.5", "lab": "Moonshot AI", "released": "2026-01-27", "openWeights": "Yes", "context": 256000, "modalities": [ "Text", "Image", "Video" ], "inputPrice": null, "outputPrice": null, "notes": "Natively multimodal K2 continuation (about 15T mixed visual and text tokens); 1T total, 32B active; modified MIT license.", "source": { "title": "Kimi K2.5: Visual Agentic Intelligence (Moonshot AI)", "url": "https://www.kimi.com/en/blog/kimi-k2-5" } }, { "model": "Qwen3.5 (Qwen3.5-397B-A17B)", "lab": "Alibaba Qwen", "released": "2026-02", "openWeights": "Yes", "context": 262144, "modalities": [ "Text", "Image", "Video" ], "inputPrice": null, "outputPrice": null, "notes": "MoE with 397B total and 17B active parameters; first Qwen open-weight flagship with native vision. Apache 2.0; hosted Qwen3.5-Plus offers 1M context.", "source": { "title": "Alibaba's Qwen 3.5 397B-A17B (VentureBeat)", "url": "https://venturebeat.com/technology/alibabas-qwen-3-5-397b-a17-beats-its-larger-trillion-parameter-model-at-a" } }, { "model": "Claude Opus 4.6", "lab": "Anthropic", "released": "2026-02-05", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": 5, "outputPrice": 25, "notes": "First Opus model with a 1M-token context window (in beta at launch)", "source": { "title": "Anthropic: Claude Opus 4.6", "url": "https://www.anthropic.com/news/claude-opus-4-6" } }, { "model": "GLM-5", "lab": "Z.ai (Zhipu)", "released": "2026-02-12", "openWeights": "Yes", "context": 202752, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "MoE with 744B total and 40B active parameters, adopting DeepSeek Sparse Attention; MIT license.", "source": { "title": "GLM-5: From Vibe Coding to Agentic Engineering (Z.ai blog)", "url": "https://z.ai/blog/glm-5" } }, { "model": "Claude Sonnet 4.6", "lab": "Anthropic", "released": "2026-02-17", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": 3, "outputPrice": 15, "notes": "1M-token context in beta at launch; default model for Free and Pro plans", "source": { "title": "Anthropic: Introducing Sonnet 4.6", "url": "https://www.anthropic.com/news/claude-sonnet-4-6" } }, { "model": "Gemini 3.1 Pro", "lab": "Google", "released": "2026-02-19", "openWeights": "No", "context": 1048576, "modalities": [ "Text", "Image", "Audio", "Video", "PDF" ], "inputPrice": 2, "outputPrice": 12, "notes": "Released in preview in the Gemini API and Gemini app; price for prompts up to 200K tokens", "source": { "title": "Google: Gemini 3.1 Pro", "url": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-pro/" } }, { "model": "GPT-5.4", "lab": "OpenAI", "released": "2026-03-05", "openWeights": "No", "context": 1050000, "modalities": [ "Text", "Image" ], "inputPrice": 2.5, "outputPrice": 15, "notes": "First GPT-5 model with a 1M-token context window and built-in computer use", "source": { "title": "OpenAI API changelog (Mar 5, 2026)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "Mistral Small 4", "lab": "Mistral AI", "released": "2026-03-16", "openWeights": "Yes", "context": 256000, "modalities": [ "Text", "Image" ], "inputPrice": 0.15, "outputPrice": 0.6, "notes": "119B total / 6.5B active; unifies instruct, reasoning and coding; Apache 2.0.", "source": { "title": "Introducing Mistral Small 4 (Mistral AI)", "url": "https://mistral.ai/news/mistral-small-4" } }, { "model": "Gemma 4", "lab": "Google", "released": "2026-04-02", "openWeights": "Yes", "context": 256000, "modalities": [ "Text", "Image", "Audio" ], "inputPrice": null, "outputPrice": null, "notes": "Released under the Apache 2.0 license; up to 256K context; audio input on smaller sizes", "source": { "title": "Google: Gemma 4", "url": "https://blog.google/innovation-and-ai/technology/developers-tools/gemma-4/" } }, { "model": "Claude Mythos Preview", "lab": "Anthropic", "released": "2026-04-07", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": 25, "outputPrice": 125, "notes": "Restricted to Project Glasswing cybersecurity partners; Anthropic does not plan general availability", "source": { "title": "Anthropic: Project Glasswing", "url": "https://www.anthropic.com/glasswing" } }, { "model": "Muse Spark", "lab": "Meta", "released": "2026-04-08", "openWeights": "No", "context": null, "modalities": [ "Text", "Image" ], "inputPrice": null, "outputPrice": null, "notes": "First model from Meta Superintelligence Labs; launched in meta.ai and the Meta AI app, API only in private preview.", "source": { "title": "Introducing Muse Spark (Meta AI)", "url": "https://ai.meta.com/blog/introducing-muse-spark-msl/" } }, { "model": "Claude Opus 4.7", "lab": "Anthropic", "released": "2026-04-16", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": 5, "outputPrice": 25, "notes": "Higher-resolution vision; introduced the current Claude tokenizer", "source": { "title": "Anthropic: Introducing Claude Opus 4.7", "url": "https://www.anthropic.com/news/claude-opus-4-7" } }, { "model": "DeepSeek-V4-Flash", "lab": "DeepSeek", "released": "2026-04-24", "openWeights": "Yes", "context": 1000000, "modalities": [ "Text" ], "inputPrice": 0.14, "outputPrice": 0.28, "notes": "Smaller V4 preview model with 284B total and 13B active parameters, MIT license.", "source": { "title": "DeepSeek V4 Preview Release (DeepSeek API news)", "url": "https://api-docs.deepseek.com/news/news260424" } }, { "model": "DeepSeek-V4-Pro", "lab": "DeepSeek", "released": "2026-04-24", "openWeights": "Yes", "context": 1000000, "modalities": [ "Text" ], "inputPrice": 1.74, "outputPrice": 3.48, "notes": "Preview release; MoE with 1.6T total and 49B active parameters, MIT license. Launched at a 75% discount; general availability 2026-08-13.", "source": { "title": "DeepSeek V4 Preview Release (DeepSeek API news)", "url": "https://api-docs.deepseek.com/news/news260424" } }, { "model": "GPT-5.5", "lab": "OpenAI", "released": "2026-04-24", "openWeights": "No", "context": 1050000, "modalities": [ "Text", "Image" ], "inputPrice": 5, "outputPrice": 30, "notes": "Released with GPT-5.5 Pro for harder problems", "source": { "title": "OpenAI API changelog (Apr 24, 2026)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "Mistral Medium 3.5", "lab": "Mistral AI", "released": "2026-04-28", "openWeights": "Yes", "context": 256000, "modalities": [ "Text", "Image" ], "inputPrice": 1.5, "outputPrice": 7.5, "notes": "Dense 128B model merging instruct, reasoning and coding; modified MIT license.", "source": { "title": "Mistral changelog (Mistral AI docs)", "url": "https://docs.mistral.ai/resources/changelogs" } }, { "model": "Command A+", "lab": "Cohere", "released": "2026-05-20", "openWeights": "Yes", "context": 128000, "modalities": [ "Text", "Image" ], "inputPrice": null, "outputPrice": null, "notes": "Cohere's first mixture-of-experts model (218B total, 25B active); Apache 2.0 license", "source": { "title": "Cohere changelog: Announcing Command A+", "url": "https://docs.cohere.com/changelog" } }, { "model": "Claude Opus 4.8", "lab": "Anthropic", "released": "2026-05-28", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": 5, "outputPrice": 25, "notes": "Fast mode priced at $10/$50", "source": { "title": "Anthropic: Introducing Claude Opus 4.8", "url": "https://www.anthropic.com/news/claude-opus-4-8" } }, { "model": "MiniMax-M3", "lab": "MiniMax", "released": "2026-06-01", "openWeights": "Yes", "context": 1000000, "modalities": [ "Text", "Image", "Video" ], "inputPrice": null, "outputPrice": null, "notes": "Native multimodal MoE (about 428B total, 23B active) with MiniMax Sparse Attention; weights followed within about 10 days under the MiniMax Community License.", "source": { "title": "MiniMax M3: Frontier Coding, 1M Context, Native Multimodality (MiniMax)", "url": "https://www.minimax.io/blog/minimax-m3" } }, { "model": "Nemotron 3 Ultra", "lab": "NVIDIA", "released": "2026-06-04", "openWeights": "Yes", "context": 1000000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "550B total / 55B active hybrid Mamba-Transformer mixture-of-experts; OpenMDW-1.1 license.", "source": { "title": "NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 model card (Hugging Face)", "url": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16" } }, { "model": "Claude Fable 5", "lab": "Anthropic", "released": "2026-06-09", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": 10, "outputPrice": 50, "notes": "New tier above Opus; Claude Mythos 5, released the same day, is offered only through a trusted access program", "source": { "title": "Anthropic: Claude Fable 5 and Claude Mythos 5", "url": "https://www.anthropic.com/news/claude-fable-5-mythos-5" } }, { "model": "GLM-5.2", "lab": "Z.ai (Zhipu)", "released": "2026-06-16", "openWeights": "Yes", "context": 1000000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Long-horizon model with a 1M-token context, MIT license.", "source": { "title": "GLM-5.2: Built for Long-Horizon Tasks (Z.ai blog)", "url": "https://z.ai/blog/glm-5.2" } }, { "model": "Claude Sonnet 5", "lab": "Anthropic", "released": "2026-06-30", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": 2, "outputPrice": 10, "notes": "Launched at $2/$10 introductory pricing, later made the standard price", "source": { "title": "Anthropic: Introducing Claude Sonnet 5", "url": "https://www.anthropic.com/news/claude-sonnet-5" } }, { "model": "GPT-5.6 Sol", "lab": "OpenAI", "released": "2026-07-09", "openWeights": "No", "context": 1050000, "modalities": [ "Text", "Image" ], "inputPrice": 4, "outputPrice": 20, "notes": "Flagship of the GPT-5.6 family, released with GPT-5.6 Terra ($2/$12) and Luna ($0.20/$1.20)", "source": { "title": "OpenAI API changelog (Jul 9, 2026)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "Inkling", "lab": "Thinking Machines Lab", "released": "2026-07-15", "openWeights": "Yes", "context": 1000000, "modalities": [ "Text", "Image", "Audio" ], "inputPrice": null, "outputPrice": null, "notes": "Thinking Machines' first model; 975B total / 41B active mixture-of-experts; Apache 2.0.", "source": { "title": "Inkling: Our Open-Weights Model (Thinking Machines Lab)", "url": "https://thinkingmachines.ai/news/introducing-inkling/" } }, { "model": "Kimi K3", "lab": "Moonshot AI", "released": "2026-07-16", "openWeights": "Yes", "context": 1048576, "modalities": [ "Text", "Image", "Video" ], "inputPrice": 3, "outputPrice": 15, "notes": "2.8T-parameter MoE (model card: 104B active) that Moonshot calls the first open 3T-class model; weights promised by 2026-07-27 under the Kimi K3 License.", "source": { "title": "Kimi K3: Open Frontier Intelligence (Moonshot AI)", "url": "https://www.kimi.com/en/blog/kimi-k3" } }, { "model": "Grok 4.5", "lab": "xAI", "released": "2026-07-16", "openWeights": "No", "context": 500000, "modalities": [ "Text", "Image" ], "inputPrice": 2, "outputPrice": 6, "notes": "Published under the SpaceXSI brand (SpaceX Super Intelligence LLC) on x.ai; aimed at coding, agentic tasks and knowledge work.", "source": { "title": "Introducing Grok 4.5 (SpaceXSI)", "url": "https://x.ai/news/grok-4-5" } }, { "model": "Claude Opus 5", "lab": "Anthropic", "released": "2026-07-24", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": 5, "outputPrice": 25, "notes": "Anthropic says it comes close to Fable 5 at half the price", "source": { "title": "Anthropic: Introducing Claude Opus 5", "url": "https://www.anthropic.com/news/claude-opus-5" } }, { "model": "Qwen3.8-Max", "lab": "Alibaba Qwen", "released": "2026-08-03", "openWeights": "Yes", "context": 1000000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "MoE with 2.4T total and 95B active parameters; API first, open weights (Qwen3.8-2.4T-A95B, text-only, custom license) followed in August 2026.", "source": { "title": "Alibaba Unveils Qwen3.8-Max (Alibaba Group)", "url": "https://www.alibabagroup.com/document-2021044032125272064" } }, { "model": "MAI-Thinking-1", "lab": "Microsoft", "released": "2026-08-12", "openWeights": "No", "context": 256000, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Microsoft AI's first reasoning model; announced at Build on 2026-06-02 in private preview, public preview on Microsoft Foundry from 2026-08-12", "source": { "title": "Microsoft AI: Introducing MAI-Thinking-1", "url": "https://microsoft.ai/news/introducing-mai-thinking-1/" } }, { "model": "GLM-5.3", "lab": "Z.ai (Zhipu)", "released": "2026-08-18", "openWeights": "Yes", "context": 1048576, "modalities": [ "Text" ], "inputPrice": null, "outputPrice": null, "notes": "Same base as GLM-5.2 with stronger coding; moved from MIT to a custom GLM-5.3 license with conditions for large model-as-a-service operators.", "source": { "title": "Z.ai release notes (New Released)", "url": "https://docs.z.ai/release-notes/new-released" } }, { "model": "Claude Fable 5.1", "lab": "Anthropic", "released": "2026-09-01", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": 10, "outputPrice": 50, "notes": "Same model as Claude Mythos 5.1 with different safeguards; Mythos 5.1 is limited to trusted access programs", "source": { "title": "Claude Docs: Claude Fable 5.1 overview", "url": "https://platform.claude.com/docs/en/models/fable-5-1/overview" } }, { "model": "Gemini 3.8 Flash", "lab": "Google", "released": "2026-09-02", "openWeights": "No", "context": 1048576, "modalities": [ "Text", "Image", "Audio", "Video", "PDF" ], "inputPrice": 0.75, "outputPrice": 3.75, "notes": "Introductory price through December 31, 2026, rising to $1.50/$7.50", "source": { "title": "Google: Gemini 3.8 Flash and 3.8 Flash Cyber", "url": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/3-8-flash-and-3-8-flash-cyber/" } }, { "model": "Muse Spark 1.3", "lab": "Meta", "released": "2026-09-02", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image", "Video", "PDF" ], "inputPrice": 1.25, "outputPrice": 4.25, "notes": "Available in Muse Code and the Meta Model API; a cheaper contributor tier lets Meta use data to improve products.", "source": { "title": "Introducing Muse Spark 1.3 (Meta AI Research)", "url": "https://research.meta.ai/blog/introducing-muse-spark-1-3" } }, { "model": "GPT-6 Astra", "lab": "OpenAI", "released": "2026-09-03", "openWeights": "No", "context": 1050000, "modalities": [ "Text", "Image" ], "inputPrice": 10, "outputPrice": 50, "notes": "OpenAI's flagship GPT-6 model; does not support the 'none' reasoning level", "source": { "title": "OpenAI API changelog (Sep 3, 2026)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "DeepSeek-V4.1-Flash", "lab": "DeepSeek", "released": "2026-09-10", "openWeights": "Yes", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": 0.3, "outputPrice": 1.2, "notes": "552B-parameter MoE with a causal encoder-decoder design (8B active for input, 16B for output) and native vision; MIT license. Prices are peak rates; off-peak is half.", "source": { "title": "DeepSeek-V4.1-Flash Release (DeepSeek API news)", "url": "https://api-docs.deepseek.com/news/news260910" } }, { "model": "Grok 4.7", "lab": "xAI", "released": "2026-09-21", "openWeights": "No", "context": 500000, "modalities": [ "Text", "Image" ], "inputPrice": 2, "outputPrice": 6, "notes": "New, larger base model than Grok 4.6, same price.", "source": { "title": "Introducing Grok 4.7 (SpaceXSI)", "url": "https://x.ai/news/grok-4-7" } }, { "model": "Claude Opus 5.5", "lab": "Anthropic", "released": "2026-09-22", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": 4, "outputPrice": 20, "notes": "First model in the Claude 5.5 family; prices 20% below Opus 5", "source": { "title": "Anthropic: Claude Opus 5.5", "url": "https://www.anthropic.com/claude-opus-5-5" } }, { "model": "GPT-6 Luna", "lab": "OpenAI", "released": "2026-09-22", "openWeights": "No", "context": 1050000, "modalities": [ "Text", "Image" ], "inputPrice": 0.1, "outputPrice": 0.5, "notes": "Lowest-cost GPT-6 model, for high-volume tasks", "source": { "title": "OpenAI API changelog (Sep 22, 2026)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "GPT-6 Sol", "lab": "OpenAI", "released": "2026-09-22", "openWeights": "No", "context": 1050000, "modalities": [ "Text", "Image" ], "inputPrice": 2, "outputPrice": 10, "notes": "Half the per-token price of GPT-5.6 Sol; replaced by GPT-6.1 Sol a week later", "source": { "title": "OpenAI API changelog (Sep 22, 2026)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "MiMo-V2.6-Pro", "lab": "Xiaomi", "released": "2026-09-22", "openWeights": "Yes", "context": 1000000, "modalities": [ "Text", "Image", "Video", "Audio" ], "inputPrice": null, "outputPrice": null, "notes": "MoE with 1.02T total and 42B active parameters; Xiaomi released weights, technical report, RL environments and code under MIT.", "source": { "title": "Introducing MiMo-V2.6 series (Xiaomi MiMo)", "url": "https://mimo.xiaomi.com/mimo-v2-6" } }, { "model": "Claude Sonnet 5.5", "lab": "Anthropic", "released": "2026-09-28", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": 2, "outputPrice": 10, "notes": "Same price as Sonnet 5 with cache reads halved to $0.10", "source": { "title": "Anthropic: Claude Sonnet 5.5", "url": "https://www.anthropic.com/claude-sonnet-5-5" } }, { "model": "GPT-6.1 Sol", "lab": "OpenAI", "released": "2026-09-29", "openWeights": "No", "context": 1050000, "modalities": [ "Text", "Image" ], "inputPrice": 2, "outputPrice": 10, "notes": "Released at DevDay 2026; same list price as GPT-6 Sol", "source": { "title": "OpenAI API changelog (Sep 29, 2026)", "url": "https://developers.openai.com/api/docs/changelog" } }, { "model": "Mistral Large 4", "lab": "Mistral AI", "released": "2026-10-06", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": null, "outputPrice": null, "notes": "52B active / 1.05T total mixture-of-experts in public preview; Mistral says open weights are coming soon. Model card lists $1.36/$4.18; a 50% launch discount applied for two weeks.", "source": { "title": "Mistral Large 4 model card (Mistral AI docs)", "url": "https://docs.mistral.ai/models/mistral-large-4-0" } }, { "model": "Claude Haiku 5.5", "lab": "Anthropic", "released": "2026-10-07", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": 0.1, "outputPrice": 0.5, "notes": "Price shown is for prompts up to 100K tokens; longer prompts cost $0.50/$2.50", "source": { "title": "Anthropic: Introducing Claude Haiku 5.5", "url": "https://www.anthropic.com/news/claude-haiku-5-5" } }, { "model": "Muse Wonder", "lab": "Meta", "released": "2026-10-12", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": 4, "outputPrice": 20, "notes": "Meta's first frontier model, for long-running agents and computer use; available via Meta Model API and Muse Code.", "source": { "title": "Muse Wonder: A Frontier Model For Personal & Professional AI Agents (Meta AI Research)", "url": "https://research.meta.ai/blog/muse-wonder" } }, { "model": "Gemini 4 Argon", "lab": "Google", "released": "2026-10-13", "openWeights": "No", "context": 1048576, "modalities": [ "Text", "Image", "Audio", "Video", "PDF" ], "inputPrice": 2, "outputPrice": 10, "notes": "Announced September 30, 2026 for Fairwind Program cyber defenders; GA October 13 at an introductory price through November 30, then $4/$20; context shared by input, thinking and output", "source": { "title": "Google: Gemini 4 Argon is now available", "url": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-argon/" } }, { "model": "Muse Spark Open", "lab": "Meta", "released": "2026-10-15", "openWeights": "Yes", "context": 1000000, "modalities": [ "Text", "Image", "Audio", "Video" ], "inputPrice": null, "outputPrice": null, "notes": "816B total / 42B active; Meta calls it the most capable US open-weight model to date. Muse Model License with limited commercial terms.", "source": { "title": "Introducing Muse Spark Open (Meta AI Research)", "url": "https://research.meta.ai/blog/muse-spark-open" } }, { "model": "Kimi K3.1 Pro", "lab": "Moonshot AI", "released": "2026-10-24", "openWeights": "No", "context": 1000000, "modalities": [ "Text", "Image" ], "inputPrice": null, "outputPrice": null, "notes": "Available in Kimi apps and API; Moonshot says full weights will be released by 2026-11-04, so weights were not yet public as of 2026-10-29.", "source": { "title": "Kimi K3.1 Pro (Moonshot AI)", "url": "https://www.kimi.com/en/blog/kimi-k3-1-pro" } }, { "model": "GPT-6.1 Terra", "lab": "OpenAI", "released": "2026-10-26", "openWeights": "No", "context": 1050000, "modalities": [ "Text", "Image" ], "inputPrice": 1, "outputPrice": 5, "notes": "Mid-priced GPT-6 model for visual reasoning and computer use", "source": { "title": "OpenAI API changelog (Oct 26, 2026)", "url": "https://developers.openai.com/api/docs/changelog" } } ] }