PulseAugur
EN
LIVE 04:33:52
中文(ZH) 美团LongCat-2.0正式发布

Meituan open-sources LongCat-2.0 trillion-parameter AI model with 1M context

Meituan has officially released its new generation trillion-parameter large model, LongCat-2.0, and made it open source. This model, trained entirely on a 50,000-card domestic computing cluster, supports a 1 million token context window and is designed for efficient and stable code understanding, generation, and execution in Agentic Coding tasks. AI

IMPACT This open-source release of a trillion-parameter model with a 1M context window could accelerate research and development in agentic coding tasks.

RANK_REASON Frontier-lab model release with system card. [lever_c_demoted from frontier_release: ic=1 ai=1.0]

Read on 36氪 (36Kr) →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Meituan open-sources LongCat-2.0 trillion-parameter AI model with 1M context

COVERAGE [1]

  1. 36氪 (36Kr) TIER_1 中文(ZH) ·

    Meituan LongCat-2.0 Officially Released

    36氪获悉,美团正式发布新一代万亿参数大模型LongCat-2.0,并将对外开源。据了解,作为业界首个在五万卡国产算力集群上完成全流程训练与推理的万亿参数模型(总参数1.6 T,平均激活约48 B,动态范围 33B~56B),LongCat-2.0从零开始预训练,原生支持1M超长上下文,其架构设计自始至终围绕一个核心目标:让模型在真实的Agentic Coding任务中,更高效、更稳定地完成代码理解、生成与执行。