Meituan launches LongCat-2.5-Preview with 1.6T parameters
美团上线 LongCat-2.5-Preview 模型：1.6T 参数，主打 AI 长程任务与多模态能力
Meituan, better known for food delivery, is now shipping frontier-scale open models with 1M-token context, a signal that Chinese tech giants outside the usual AI labs are competing on agentic long-horizon workloads. Developers testing coding agents can plug it into Claude Code or Kilo Code directly and compare it against DeepSeek or Qwen on the same workflows. Availability via API and web makes hands-on trials straightforward.
Meituan's LongCat API open platform released LongCat-2.
5-Preview on 25 Sep 2025.The MoE model has about 1.
6 trillion total parameters and activates roughly 48 billion per inference.It natively supports a 1 million token context window for long documents, codebases, logs and multi-turn tasks.
New features include image understanding for cross-modal QA, summarisation and visual reasoning, plus coding ability.
It works with Claude Code, Hermes, OpenClaw, OpenCode and Kilo Code, and is available via API and web.
美团上线 LongCat-2.5-Preview 模型：1.6T 参数，主打 AI 长程任务与多模态能力
IT之家 9 月 26 日消息，美团旗下 LongCat API 开放平台于 9 月 25 日上线新一代大模型 LongCat-2.5-Preview， 主打“长程任务”与多模态能力，并同步开放 API 与网页端体验入口。 IT之家附上官方更新日志内容如下： LongCat-2.5-Preview 的核心特性如下： 多模态理解：新增图片理解能力 ，可解析图像内容，支持跨模态问答、内容摘要与复杂视觉推理。 卓越的 Coding 能力 ：在代码生成、代码理解与自动化编程任务上表现突出。 深度适配 Claude Code 等主流开发环境 ：与 Claude Code、Hermes、OpenClaw、OpenCode、Kilo Code 高效协同。 根据官方公布的信息，LongCat-2.5-Preview 模型延续 MoE（混合专家）路线，总参数量约 1.6 万亿，每次推理激活参数约 480 亿。该模型支持原生支持 100 万 token 上下文窗口，适合处理超长文档、代码库、日志与多轮复杂任务。 参考 LongCat API 开放平台更新日志
This report is categorized under AITOP's "Models" desk, focusing on China's AI industry and domestic foundation model ecosystem. Within our cross-regional analytical framework, it directly maps to US-China foundation model benchmarks and commercialization comparison and AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include LongCat, Meituan, MoE, multimodal.
Source attribution: curated from IT之家. Cross-source corroboration: 11 independent sources have covered this development (including pandaily, berryxia, Garry Tan), confirming industry-wide signal strength.
