{"version":1,"generatedAt":"2026-11-10T01:56:29.783Z","attribution":"Compiled by HARBAN AI LABS from the linked sources. Cite the original source for any number you reuse.","documentation":"https://harban-ai-labs.vercel.app/benchmarks/methodology","count":14,"data":[{"slug":"claude-opus-5-5","url":"https://harban-ai-labs.vercel.app/models/claude-opus-5-5","name":"Claude Opus 5.5","organization":{"id":"anthropic","name":"Anthropic","url":"https://www.anthropic.com"},"family":"Claude 5.5","status":"current","releaseDate":"2026-09-22","releaseDateKind":"released","contextWindow":1000000,"contextWindowAsStated":"1M tokens","maxOutputTokens":128000,"maxOutputTokensAsStated":"128K tokens (synchronous Messages API; up to 300K on Message Batches API with beta header output-300k-2026-03-24)","knowledgeCutoff":"2026-06 (reliable knowledge cutoff)","inputModalities":["text","image"],"outputModalities":["text"],"openWeights":false,"license":null,"apiModelId":"claude-opus-5-5","links":{"announcement":"https://www.anthropic.com/claude-opus-5-5","docs":"https://platform.claude.com/docs/en/models/opus-5-5/overview","modelCard":"https://www.anthropic.com/claude-opus-5-5-system-card"},"sources":{"releaseDate":"https://platform.claude.com/docs/en/models/opus-5-5/overview","contextWindow":"https://platform.claude.com/docs/en/models/opus-5-5/overview","maxOutputTokens":"https://platform.claude.com/docs/en/models/opus-5-5/overview","inputModalities":"https://platform.claude.com/docs/en/models/opus-5-5/overview","outputModalities":"https://platform.claude.com/docs/en/models/opus-5-5/overview","apiModelId":"https://platform.claude.com/docs/en/models/overview","knowledgeCutoff":"https://platform.claude.com/docs/en/models/overview"},"verifiedOn":"2026-09-28","notes":"Release date: labelled “Released” in Anthropic’s model documentation (22 Sep 2026), the same day as the announcement."},{"slug":"grok-4-7","url":"https://harban-ai-labs.vercel.app/models/grok-4-7","name":"Grok 4.7","organization":{"id":"xai","name":"xAI","url":"https://x.ai"},"family":"Grok 4","status":"current","releaseDate":"2026-09-21","releaseDateKind":"announced","contextWindow":500000,"contextWindowAsStated":"500,000 tokens","maxOutputTokens":null,"maxOutputTokensAsStated":"No text output limit","knowledgeCutoff":"May 2026 per xAI API docs; the model card states a June 2026 pretraining cutoff","inputModalities":["text","image"],"outputModalities":["text"],"openWeights":false,"license":null,"apiModelId":"grok-4.7","links":{"announcement":"https://x.ai/news/grok-4-7","docs":"https://docs.x.ai/developers/grok-4-7","modelCard":"https://media.x.ai/v1/website/4p7card-5eccc980.pdf"},"sources":{"releaseDate":"https://x.ai/news/grok-4-7","contextWindow":"https://docs.x.ai/developers/grok-4-7","inputModalities":"https://docs.x.ai/developers/grok-4-7","outputModalities":"https://docs.x.ai/developers/grok-4-7","apiModelId":"https://docs.x.ai/developers/release-notes","knowledgeCutoff":"https://docs.x.ai/developers/grok-4-7","maxOutputTokens":"https://docs.x.ai/developers/grok-4-7"},"verifiedOn":"2026-09-28","notes":"Release date: announcement date on x.ai. Reasoning levels: low, medium, high (default) and xhigh (per docs). xAI pages brand the company \"SpaceXAI\". No official open-weights release was found; the model is served through xAI’s API."},{"slug":"deepseek-v4-1-flash","url":"https://harban-ai-labs.vercel.app/models/deepseek-v4-1-flash","name":"DeepSeek-V4.1-Flash","organization":{"id":"deepseek","name":"DeepSeek","url":"https://www.deepseek.com"},"family":"DeepSeek-V4","status":"current","releaseDate":"2026-09-10","releaseDateKind":"announced","contextWindow":1000000,"contextWindowAsStated":"1M (API docs); 'up to one million tokens' (HF card)","maxOutputTokens":384000,"maxOutputTokensAsStated":"MAXIMUM: 384K","knowledgeCutoff":null,"inputModalities":["text","image"],"outputModalities":["text"],"openWeights":true,"license":{"name":"MIT License","url":"https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash/blob/main/LICENSE"},"apiModelId":"deepseek-flash","links":{"announcement":"https://api-docs.deepseek.com/news/news260910","docs":"https://api-docs.deepseek.com/quick_start/pricing","modelCard":"https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash"},"sources":{"releaseDate":"https://api-docs.deepseek.com/news/news260910","contextWindow":"https://api-docs.deepseek.com/quick_start/pricing","maxOutputTokens":"https://api-docs.deepseek.com/quick_start/pricing","inputModalities":"https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash","outputModalities":"https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash","openWeights":"https://api-docs.deepseek.com/news/news260910","apiModelId":"https://api-docs.deepseek.com/quick_start/pricing","license":"https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash"},"verifiedOn":"2026-09-28","notes":"Release date: announcement, API and weights release. Per the model card: a 552B-parameter mixture-of-experts model (8B parameters active for prefill, 16B for decode)."},{"slug":"gpt-6-astra","url":"https://harban-ai-labs.vercel.app/models/gpt-6-astra","name":"GPT-6 Astra","organization":{"id":"openai","name":"OpenAI","url":"https://openai.com"},"family":"GPT-6","status":"current","releaseDate":"2026-09-03","releaseDateKind":"released","contextWindow":1050000,"contextWindowAsStated":"1,050,000 context window","maxOutputTokens":128000,"maxOutputTokensAsStated":"128,000 max output tokens","knowledgeCutoff":"2026-04-30","inputModalities":["text","image"],"outputModalities":["text"],"openWeights":false,"license":null,"apiModelId":"gpt-6-astra","links":{"announcement":null,"docs":"https://developers.openai.com/api/docs/models/gpt-6-astra","modelCard":"https://deploymentsafety.openai.com/gpt-6-astra"},"sources":{"releaseDate":"https://developers.openai.com/api/docs/changelog","contextWindow":"https://developers.openai.com/api/docs/models/gpt-6-astra","maxOutputTokens":"https://developers.openai.com/api/docs/models/gpt-6-astra","inputModalities":"https://developers.openai.com/api/docs/models/gpt-6-astra","outputModalities":"https://developers.openai.com/api/docs/models/gpt-6-astra","apiModelId":"https://developers.openai.com/api/docs/models/gpt-6-astra","knowledgeCutoff":"https://developers.openai.com/api/docs/models/gpt-6-astra"},"verifiedOn":"2026-09-28","notes":"Release date: API release per OpenAI’s developer changelog (3 Sep 2026). OpenAI’s system card for this model reports safety and Preparedness evaluations — including Preparedness capability assessments such as ExploitBench, SRE-Bench and MLE-Bench Revised — rather than general capability benchmark scores such as GPQA or SWE-bench. OpenAI’s announcement page blocks automated access, so no capability scores are recorded for this model."},{"slug":"gemini-3-8-flash","url":"https://harban-ai-labs.vercel.app/models/gemini-3-8-flash","name":"Gemini 3.8 Flash","organization":{"id":"google-deepmind","name":"Google DeepMind","url":"https://deepmind.google"},"family":"Gemini 3","status":"current","releaseDate":"2026-09-02","releaseDateKind":"ga","contextWindow":1048576,"contextWindowAsStated":"Input token limit 1,048,576","maxOutputTokens":65536,"maxOutputTokensAsStated":"Output token limit 65,536","knowledgeCutoff":"March 2026 (model card; for some domains, knowledge is limited to January 2025)","inputModalities":["text","image","video","audio","pdf"],"outputModalities":["text"],"openWeights":false,"license":null,"apiModelId":"gemini-3.8-flash","links":{"announcement":"https://blog.google/innovation-and-ai/models-and-research/gemini-models/3-8-flash-and-3-8-flash-cyber/","docs":"https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash","modelCard":"https://deepmind.google/models/model-cards/gemini-3-8-flash/"},"sources":{"releaseDate":"https://ai.google.dev/gemini-api/docs/changelog","contextWindow":"https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash","maxOutputTokens":"https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash","inputModalities":"https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash","outputModalities":"https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash","apiModelId":"https://ai.google.dev/gemini-api/docs/models/gemini-3.8-flash","knowledgeCutoff":"https://deepmind.google/models/model-cards/gemini-3-8-flash/"},"verifiedOn":"2026-09-28","notes":"Release date: general availability in the Gemini API changelog."},{"slug":"qwen3-8-27b","url":"https://harban-ai-labs.vercel.app/models/qwen3-8-27b","name":"Qwen3.8-27B","organization":{"id":"alibaba-qwen","name":"Alibaba Cloud (Qwen)","url":"https://qwen.ai"},"family":"Qwen3.8","status":"current","releaseDate":"2026-08-14","releaseDateKind":"released","contextWindow":262144,"contextWindowAsStated":"262,144 natively and extensible up to 1,000,000 tokens","maxOutputTokens":131072,"maxOutputTokensAsStated":"Max Output Length 131072 (Model Studio, API)","knowledgeCutoff":null,"inputModalities":["text","image","video"],"outputModalities":["text"],"openWeights":true,"license":{"name":"Apache 2.0","url":"https://huggingface.co/Qwen/Qwen3.8-27B/blob/main/LICENSE"},"apiModelId":"qwen3.8-27b","links":{"announcement":"https://github.com/QwenLM/Qwen3.8","docs":"https://www.alibabacloud.com/help/en/model-studio/newly-released-models","modelCard":"https://huggingface.co/Qwen/Qwen3.8-27B"},"sources":{"releaseDate":"https://raw.githubusercontent.com/QwenLM/Qwen3.8/main/README.md","contextWindow":"https://huggingface.co/Qwen/Qwen3.8-27B","inputModalities":"https://huggingface.co/Qwen/Qwen3.8-27B","outputModalities":"https://huggingface.co/Qwen/Qwen3.8-27B","openWeights":"https://huggingface.co/Qwen/Qwen3.8-27B","apiModelId":"https://www.alibabacloud.com/help/en/model-studio/newly-released-models","license":"https://huggingface.co/Qwen/Qwen3.8-27B","maxOutputTokens":"https://www.alibabacloud.com/help/en/model-studio/qwen3-8-27b"},"verifiedOn":"2026-09-28","notes":"Release date: weights release per the official GitHub README (Model Studio lists the API model on 17 Aug 2026). Dense 27B vision-language model; thinking is on by default and can be disabled (model card). Two deployments are mixed in this record: the context window shown (262,144) is the open-weights native length from the model card, while the max output shown (131,072) and the API model ID belong to the hosted qwen3.8-27b model on Model Studio, which lists a 1,000,000-token context window."},{"slug":"deepseek-v4-pro","url":"https://harban-ai-labs.vercel.app/models/deepseek-v4-pro","name":"DeepSeek-V4-Pro","organization":{"id":"deepseek","name":"DeepSeek","url":"https://www.deepseek.com"},"family":"DeepSeek-V4","status":"current","releaseDate":"2026-08-13","releaseDateKind":"ga","contextWindow":1000000,"contextWindowAsStated":"1M","maxOutputTokens":384000,"maxOutputTokensAsStated":"MAXIMUM: 384K","knowledgeCutoff":null,"inputModalities":["text"],"outputModalities":null,"openWeights":true,"license":{"name":"MIT License","url":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro-0813/blob/main/LICENSE"},"apiModelId":"deepseek-v4-pro","links":{"announcement":"https://api-docs.deepseek.com/news/news260813","docs":"https://api-docs.deepseek.com/quick_start/pricing","modelCard":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro-0813"},"sources":{"releaseDate":"https://api-docs.deepseek.com/news/news260813","contextWindow":"https://api-docs.deepseek.com/quick_start/pricing","maxOutputTokens":"https://api-docs.deepseek.com/quick_start/pricing","inputModalities":"https://api-docs.deepseek.com/quick_start/pricing","openWeights":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro-0813","apiModelId":"https://api-docs.deepseek.com/quick_start/pricing","license":"https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro-0813"},"verifiedOn":"2026-09-28","notes":"Release date: general availability of DeepSeek-V4-Pro-0813 per DeepSeek’s API news (a preview was released on 24 Apr 2026). DeepSeek’s API documentation lists vision as not supported, so input is recorded as text only. Supports thinking and non-thinking modes."},{"slug":"muse-glimmer-30b","url":"https://harban-ai-labs.vercel.app/models/muse-glimmer-30b","name":"Muse Glimmer","organization":{"id":"meta","name":"Meta","url":"https://ai.meta.com"},"family":"Muse","status":"current","releaseDate":"2026-08-10","releaseDateKind":"announced","contextWindow":131072,"contextWindowAsStated":"131,072+ (HF card); '128K tokens by default; longer contexts supported' (Meta dev docs)","maxOutputTokens":null,"maxOutputTokensAsStated":null,"knowledgeCutoff":"January 4, 2026 (model card)","inputModalities":["text","image"],"outputModalities":["text"],"openWeights":true,"license":{"name":"Apache 2.0","url":"https://huggingface.co/meta-models/Muse-Glimmer-30B/blob/main/LICENSE"},"apiModelId":null,"links":{"announcement":"https://research.meta.ai/blog/introducing-muse-glimmer-open-agentic-model","docs":"https://dev.meta.ai/docs/muse-glimmer","modelCard":"https://huggingface.co/meta-models/Muse-Glimmer-30B"},"sources":{"releaseDate":"https://research.meta.ai/blog/introducing-muse-glimmer-open-agentic-model","contextWindow":"https://huggingface.co/meta-models/Muse-Glimmer-30B","inputModalities":"https://huggingface.co/meta-models/Muse-Glimmer-30B","outputModalities":"https://huggingface.co/meta-models/Muse-Glimmer-30B","openWeights":"https://research.meta.ai/blog/introducing-muse-glimmer-open-agentic-model","license":"https://huggingface.co/meta-models/Muse-Glimmer-30B","knowledgeCutoff":"https://huggingface.co/meta-models/Muse-Glimmer-30B"},"verifiedOn":"2026-09-28","notes":"Release date: announcement and open-weights release on Meta’s research blog. Published on Hugging Face under the meta-models organisation, which Meta’s blog links as official. Per the model card, audio is not supported and video is processed as frames."},{"slug":"qwen3-8-max","url":"https://harban-ai-labs.vercel.app/models/qwen3-8-max","name":"Qwen3.8-Max","organization":{"id":"alibaba-qwen","name":"Alibaba Cloud (Qwen)","url":"https://qwen.ai"},"family":"Qwen3.8","status":"current","releaseDate":"2026-08-03","releaseDateKind":"announced","contextWindow":1000000,"contextWindowAsStated":"Context Window 1000000 (max input 991808; 983616 in thinking mode)","maxOutputTokens":131072,"maxOutputTokensAsStated":"Max Output Length 131072","knowledgeCutoff":null,"inputModalities":["text","image","video"],"outputModalities":["text"],"openWeights":false,"license":null,"apiModelId":"qwen3.8-max","links":{"announcement":"https://www.alibabacloud.com/en/press-room/alibaba-unveils-qwen3-8-max","docs":"https://www.alibabacloud.com/help/en/model-studio/qwen3-8-max","modelCard":"https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B"},"sources":{"releaseDate":"https://www.alibabacloud.com/en/press-room/alibaba-unveils-qwen3-8-max","contextWindow":"https://www.alibabacloud.com/help/en/model-studio/qwen3-8-max","maxOutputTokens":"https://www.alibabacloud.com/help/en/model-studio/qwen3-8-max","inputModalities":"https://www.alibabacloud.com/help/en/model-studio/qwen3-8-max","outputModalities":"https://www.alibabacloud.com/help/en/model-studio/qwen3-8-max","apiModelId":"https://www.alibabacloud.com/help/en/model-studio/qwen3-8-max"},"verifiedOn":"2026-09-28","notes":"Release date: announcement and API availability (Alibaba Cloud press release). 2.4T total / 95B active parameters (press release). This entry describes the hosted API model; Alibaba also published related open weights as Qwen3.8-2.4T-A95B under a custom licence (see the model card link)."},{"slug":"mistral-medium-3-5","url":"https://harban-ai-labs.vercel.app/models/mistral-medium-3-5","name":"Mistral Medium 3.5","organization":{"id":"mistral-ai","name":"Mistral AI","url":"https://mistral.ai"},"family":"Mistral Medium","status":"current","releaseDate":"2026-04-28","releaseDateKind":"released","contextWindow":262144,"contextWindowAsStated":"256k (docs and model card)","maxOutputTokens":null,"maxOutputTokensAsStated":null,"knowledgeCutoff":null,"inputModalities":["text","image"],"outputModalities":["text"],"openWeights":true,"license":{"name":"Modified MIT License","url":"https://huggingface.co/mistralai/Mistral-Medium-3.5-128B/blob/main/LICENSE"},"apiModelId":"mistral-medium-3-5","links":{"announcement":"https://mistral.ai/news/vibe-remote-agents-mistral-medium-3-5","docs":"https://docs.mistral.ai/models/mistral-medium-3-5-26-04","modelCard":"https://huggingface.co/mistralai/Mistral-Medium-3.5-128B"},"sources":{"releaseDate":"https://docs.mistral.ai/resources/changelogs","contextWindow":"https://huggingface.co/mistralai/Mistral-Medium-3.5-128B/raw/main/config.json","inputModalities":"https://huggingface.co/mistralai/Mistral-Medium-3.5-128B","outputModalities":"https://huggingface.co/mistralai/Mistral-Medium-3.5-128B","openWeights":"https://docs.mistral.ai/models/mistral-medium-3-5-26-04","apiModelId":"https://docs.mistral.ai/resources/changelogs","license":"https://huggingface.co/mistralai/Mistral-Medium-3.5-128B"},"verifiedOn":"2026-09-28","notes":"Release date: API and weights release per Mistral’s documentation changelog (the blog announcement is dated 22 May 2026). A dense 128B-parameter model, version 26.04. Context window: stated as “256k”; the number shown is max_position_embeddings from the official config.json."},{"slug":"gemma-4-31b","url":"https://harban-ai-labs.vercel.app/models/gemma-4-31b","name":"Gemma 4 31B","organization":{"id":"google-deepmind","name":"Google DeepMind","url":"https://deepmind.google"},"family":"Gemma 4","status":"current","releaseDate":"2026-03-31","releaseDateKind":"released","contextWindow":262144,"contextWindowAsStated":"256K tokens (model card)","maxOutputTokens":null,"maxOutputTokensAsStated":null,"knowledgeCutoff":"January 2025 (pre-training data cutoff, per model card)","inputModalities":["text","image"],"outputModalities":["text"],"openWeights":true,"license":{"name":"Apache 2.0","url":"https://ai.google.dev/gemma/apache_2"},"apiModelId":"gemma-4-31b-it","links":{"announcement":"https://blog.google/innovation-and-ai/technology/developers-tools/gemma-4/","docs":"https://ai.google.dev/gemma/docs/core","modelCard":"https://ai.google.dev/gemma/docs/core/model_card_4"},"sources":{"releaseDate":"https://ai.google.dev/gemma/docs/releases","contextWindow":"https://huggingface.co/google/gemma-4-31B-it/raw/main/config.json","inputModalities":"https://huggingface.co/google/gemma-4-31B-it","outputModalities":"https://huggingface.co/google/gemma-4-31B-it","openWeights":"https://huggingface.co/google/gemma-4-31B-it","license":"https://huggingface.co/google/gemma-4-31B-it","apiModelId":"https://ai.google.dev/gemma/docs/core/gemma_on_gemini_api","knowledgeCutoff":"https://huggingface.co/google/gemma-4-31B-it"},"verifiedOn":"2026-09-28","notes":"Release date: weights release per the Gemma release notes (launch blog dated 2 Apr 2026). The 31B model takes text and image input. Context window: the model card states \"256K tokens\"; the number shown is max_position_embeddings from the official config.json. Also available through the Gemini API as gemma-4-31b-it."},{"slug":"gemini-3-1-pro","url":"https://harban-ai-labs.vercel.app/models/gemini-3-1-pro","name":"Gemini 3.1 Pro","organization":{"id":"google-deepmind","name":"Google DeepMind","url":"https://deepmind.google"},"family":"Gemini 3","status":"preview","releaseDate":"2026-02-19","releaseDateKind":"released","contextWindow":1048576,"contextWindowAsStated":"Input token limit 1,048,576","maxOutputTokens":65536,"maxOutputTokensAsStated":"Output token limit 65,536","knowledgeCutoff":null,"inputModalities":["text","image","video","audio","pdf"],"outputModalities":["text"],"openWeights":false,"license":null,"apiModelId":"gemini-3.1-pro-preview","links":{"announcement":"https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-pro/","docs":"https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview","modelCard":"https://deepmind.google/models/model-cards/gemini-3-1-pro/"},"sources":{"releaseDate":"https://ai.google.dev/gemini-api/docs/changelog","contextWindow":"https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview","maxOutputTokens":"https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview","inputModalities":"https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview","outputModalities":"https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview","apiModelId":"https://ai.google.dev/gemini-api/docs/models/gemini-3.1-pro-preview"},"verifiedOn":"2026-09-28","notes":"Release date: preview release in the Gemini API changelog. Labelled “Preview” in Google’s documentation as of 28 Sep 2026."},{"slug":"gpt-oss-120b","url":"https://harban-ai-labs.vercel.app/models/gpt-oss-120b","name":"gpt-oss-120b","organization":{"id":"openai","name":"OpenAI","url":"https://openai.com"},"family":"gpt-oss","status":"current","releaseDate":"2025-08-05","releaseDateKind":"announced","contextWindow":131072,"contextWindowAsStated":"131,072 context window","maxOutputTokens":131072,"maxOutputTokensAsStated":"131,072 max output tokens","knowledgeCutoff":"2024-06-01","inputModalities":["text"],"outputModalities":["text"],"openWeights":true,"license":{"name":"Apache 2.0","url":"https://huggingface.co/openai/gpt-oss-120b/blob/main/LICENSE"},"apiModelId":"gpt-oss-120b","links":{"announcement":"https://openai.com/index/introducing-gpt-oss/","docs":"https://developers.openai.com/api/docs/models/gpt-oss-120b","modelCard":"https://cdn.openai.com/pdf/419b6906-9da6-406c-a19d-1bb078ac7637/oai_gpt-oss_model_card.pdf"},"sources":{"releaseDate":"https://cdn.openai.com/pdf/419b6906-9da6-406c-a19d-1bb078ac7637/oai_gpt-oss_model_card.pdf","contextWindow":"https://developers.openai.com/api/docs/models/gpt-oss-120b","maxOutputTokens":"https://developers.openai.com/api/docs/models/gpt-oss-120b","inputModalities":"https://developers.openai.com/api/docs/models/gpt-oss-120b","outputModalities":"https://developers.openai.com/api/docs/models/gpt-oss-120b","openWeights":"https://huggingface.co/openai/gpt-oss-120b","apiModelId":"https://developers.openai.com/api/docs/models/gpt-oss-120b","knowledgeCutoff":"https://developers.openai.com/api/docs/models/gpt-oss-120b","license":"https://huggingface.co/openai/gpt-oss-120b"},"verifiedOn":"2026-09-28","notes":"Release date from the official model card. 117B total / 5.1B active parameters (OpenAI model page). OpenAI’s API docs list the ID gpt-oss-120b, but with a rate limit of 0 on every usage tier (“Not supported” on the free tier), so it is not presented here as callable on OpenAI’s own API; the weights are openly available."},{"slug":"llama-4-maverick","url":"https://harban-ai-labs.vercel.app/models/llama-4-maverick","name":"Llama 4 Maverick","organization":{"id":"meta","name":"Meta","url":"https://ai.meta.com"},"family":"Llama 4","status":"current","releaseDate":"2025-04-05","releaseDateKind":"released","contextWindow":1000000,"contextWindowAsStated":"1M","maxOutputTokens":null,"maxOutputTokensAsStated":null,"knowledgeCutoff":"2024-08","inputModalities":["text","image"],"outputModalities":["text"],"openWeights":true,"license":{"name":"Llama 4 Community License Agreement","url":"https://github.com/meta-llama/llama-models/blob/main/models/llama4/LICENSE"},"apiModelId":null,"links":{"announcement":"https://ai.meta.com/blog/llama-4-multimodal-intelligence/","docs":"https://dev.meta.ai/llama/docs/model-cards-and-prompt-formats/llama4","modelCard":"https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct"},"sources":{"releaseDate":"https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct","contextWindow":"https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct","inputModalities":"https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct","outputModalities":"https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct","knowledgeCutoff":"https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct","license":"https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct","openWeights":"https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct"},"verifiedOn":"2026-09-28","notes":"Release date: \"Model Release Date\" on the official model card. 17B active / 400B total parameters (model card)."}]}