{"id":"longcat-2-5","slug":"longcat-2-5","name":"LongCat-2.5","orgId":"meituan","family":"longcat-2-5","releaseDate":null,"availability":"announced","openWeights":false,"license":null,"paramsTotal":1600000000000,"paramsSource":"tracked","contextWindow":null,"knowledgeCutoff":null,"modelCardUrl":"https://longcat.ai/blog/longcat-2.5/","hfRepo":null,"modalities":["text","image","video"],"aliases":["LongCat-2.5","longcat-2.5","longcat 2.5","longcat2.5","meituan longcat-2.5","meituan-longcat/longcat-2.5"],"aliasNote":null,"origin":"tracked","tracked":{"trackedSince":"2026-11-02","announced":"2026-11-01","blurb":"Meituan LongCat team's native multimodal model for long-horizon agentic work such as coding, browsing and tool use: a 1.6-trillion-parameter mixture-of-experts model with about 48B parameters active per token, building on LongCat-2.0. It takes text, images and video, has a 1M-token context and up to 256K output tokens. Meituan says training ran entirely on domestic Chinese AI accelerators. Available in LongCat chat and the LongCat API (OpenAI- and Anthropic-compatible); weights are listed as coming soon.","api":"live","apiModelId":"LongCat-2.5","openWeights":"announced","params":{"text":"1.6T total, about 48B active per token (mixture of experts)","source":"LongCat-2.5 launch blog (longcat.ai)","total":1600000000000},"facts":[{"label":"Type","value":"Native multimodal model for long-horizon agentic tasks: text, image and video in, text out; thinking can be switched on or off"},{"label":"Architecture","value":"Mixture of experts, 1.6T total and about 48B active per token; visual input is represented as multi-level discrete tokens"},{"label":"Context","value":"1,048,576 tokens (1M), up to 262,144 output tokens (256K), per the LongCat API docs"},{"label":"Training","value":"Builds on LongCat-2.0 with long-context training, large-scale long-horizon agentic reinforcement learning and multi-teacher distillation; trained entirely on domestic Chinese AI accelerators, per Meituan"},{"label":"Inference","value":"W4A4 quantized serving on domestic AI accelerators, per the LongCat blog"},{"label":"API","value":"LongCat API Platform, model LongCat-2.5, at api.longcat.chat with OpenAI Chat Completions, OpenAI Responses and Anthropic Messages compatible endpoints; listed as available from 2026-11-02 in the API changelog"},{"label":"Price","value":"$0.30 per 1M input tokens, $0.006 per 1M cached input tokens and $1.20 per 1M output tokens (¥2, ¥0.04 and ¥8), labelled a limited-time discounted price by LongCat; prices are subject to change"},{"label":"Availability","value":"LongCat chat at longcat.ai and the LongCat API"},{"label":"Weights","value":"Not yet published: the blog lists GitHub and Hugging Face as coming soon, and the meituan-longcat Hugging Face org had no LongCat-2.5 repo as of Nov 2, 2026"},{"label":"Earlier preview","value":"LongCat-2.5-Preview launched on the LongCat API on Sep 25, 2026 (image understanding and coding), per the API changelog; a separate earlier release, not tracked here"},{"label":"Predecessor","value":"LongCat-2.0, released on the LongCat API on Jun 30, 2026, per the API changelog"}],"claims":["Lab claim: 88.4 on Terminal Bench v2.1, 38.5 on Terminal Bench v4.0, 64.6 on DeepSWE v1.1, 60.5 on NL2Repo and 82.1 on PaperBench in LongCat's own evaluation.","Lab claim: 91.9 on BrowseComp, 63.1 on HLE with tools (text-only subset), 83.1 on MCP Atlas and 51.8 on AutomationBench v1.0.6.","Lab claim: 78.6 on Vision2Web and 85.2 on VideoMME.","Lab claim: Meituan says its scores on several mainstream benchmarks fall between Claude-Opus-4.8 and Claude-Opus-5.","Lab claim: W4A4 serving raises prefill throughput by 43.3% and decode throughput by 42.77% and cuts TTFT by 15.91% and TPOT by 19.72% against a W8A8 INT8 baseline."],"sources":[{"kind":"lab","title":"Introducing LongCat-2.5 (LongCat blog)","url":"https://longcat.ai/blog/longcat-2.5/","date":"2026-11-01"},{"kind":"lab","title":"LongCat API changelog","url":"https://longcat.chat/platform/docs/change-log","date":"2026-11-02"},{"kind":"lab","title":"LongCat-2.5 pricing (LongCat API docs)","url":"https://longcat.chat/platform/docs/pricing/longcat-2.5"},{"kind":"lab","title":"Meituan news: LongCat-2.5 release","url":"https://www.meituan.com/news/NN261102280001343","date":"2026-11-02"},{"kind":"press","title":"aitrackerbot post on X","url":"https://x.com/aitrackerbot/status/2116967358602301733","date":"2026-11-01"}],"firstBenchmark":null},"organization":{"id":"meituan","name":"Meituan","slug":"meituan","country":null,"url":"https://longcat.ai","color":"#A39E8F","mark":"ME"},"price":null,"answer":{"word":"Pending benchmarks.","scope":"Announced Nov 1, 2026; no independent results yet.","lead":"As of Nov 3, 2026, LongCat-2.5 is pending benchmarks on isitsota: announced Nov 1, 2026, with no independent results on the boards we track yet.","confidence":"low"},"status":"PENDING_BENCHMARKS","verdicts":[]}