ModelsMulti-source72°9月26日 07:48

美团发布 LongCat 龙猫-2.5-Preview：16T 参数、1M 上下文、原生多模态

berryxia@berryxia

Post· 3 sources

View the post

Why we picked it

美团中秋半夜甩出 LongCat 2.5 预览版，16T 参数只激活 48，还有 1M 上下文和原生多模态，参数配置看着很猛，可以蹲一下实测成绩。

美团在中秋假期发布 LongCat 龙猫-2.5-Preview。

Body

该模型总参数 16T，激活参数 48，支持 1M 上下文，且为原生多模态架构。采用 MoE 路由方式，推理成本对应激活参数规模。

实际基准表现尚未在公告中给出。

Original · berryxia

尼玛，开水团中秋不休息啊… 直接这个时间发了新模型…

LongCat 龙猫-2.5-Preview 16Tb 参数 48 激活 1M 上下文原生多模态

Buff叠满，不知实际效果如何～。

Industry Coordinates & Source Verification

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, 美团, MoE, 多模态.

Source attribution: curated from berryxia (@berryxia). Cross-source corroboration: 3 independent sources have covered this development (including IT之家, Teortaxes), confirming industry-wide signal strength.

IT之家09-26 06:16Source
Teortaxes09-25 16:24Source

#LongCat #美团 #MoE #多模态 #开源模型

Keep reading

Recent stories with the same tag or section

Papers US vs China11:55

Chinese content研究表明 LLM 判断新闻标题会受框架措辞影响

LLM 判新闻会被标题措辞带偏，情绪化标题能让严重度评分高出近 1 分，DeepSeek 和 LongCat 四个模型都中招。做内容审核或新闻路由的团队该看看：补上正文能消掉大半偏差。

#DeepSeek #LongCat #MiMo #LLM 评估

arXiv: DeepSeek@Mo Zhou, Yiwei ZhangSource

Models US vs ChinaMulti-source71°00:08

Chinese content美团发布 LongCat-2.5 多模态模型：1.6T 总参数、49B 激活

美团出了个新模型 LongCat-2.5，1.6T 参数但每次只用激活 49B，官方说是直接对标 GPT-5.5 和 Claude Opus，API 已经能调了。

#LongCat #Meituan #GPT-5.5 #Claude

Event

TestingCatalog@testingcatalogSource

Models US vs ChinaOfficial10:19

Chinese content美团发布 LongCat-2.5：1.6T 原生多模态模型，面向长程智能体编程

美团放出 1.6T 的 LongCat-2.5，只激活 48B，专门练了长程智能体编程，还支持 W4A4 跑国产芯片。

#LongCat #美团 #多模态 #智能体

Event

Primary sourcePandaily@[email protected] (Pandaily)Source

Models US vs ChinaMulti-source09:21

Chinese content美团龙猫发布 LongCat-2.5：1.6T 参数原生多模态模型，主打长程智能体任务

美团把 LongCat-2.0 升级成 1.6T 参数的多模态版本，专门练长程编程和 Agent 任务，视觉和代码能力都拉满，可以在线直接试。

#LongCat #美团 #多模态 #智能体

Event

IT之家Source

Models US vs ChinaOfficialSelected73°12:13

Meituan Ships LongCat-2.5-Preview: 1.6T MoE With 1M Context

A food-delivery giant shipping a frontier-scale agent model shows how far Chinese consumer tech firms now push into foundation models, though the lack of published benchmarks makes real capability hard to judge. Developers building long-context or agentic workflows should track when independent evals and API access appear, since LongCat competes with open-weight MoE releases from Alibaba's Qwen and DeepSeek. Its 1M-token window targets codebase-scale agent tasks directly.

#LongCat #Meituan #MoE #open-source models

Event

Primary sourcepandaily@[email protected] (Pandaily)Source

Models US vs ChinaMulti-sourceSelected73°14:21

Meituan launches LongCat-2.5-Preview with 1.6T parameters

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.

#LongCat #Meituan #MoE #multimodal

Event

IT之家Source