Meituan releases a 1.6-trillion-parameter agent model that never touched an Nvidia GPU

What happened

Meituan has released the weights for LongCat-2.5, a 1.6-trillion-parameter model built for AI agents (AI that carries out multi-step tasks rather than answering one question) and trained entirely on Chinese hardware. NVIDIA is a semiconductor company based in Santa Clara, and its products and services include software.

It was trained entirely on Chinese hardware and performs at the level US models like Claude Opus 5 reached a few months ago on coding benchmarks. The LongCat team at Chinese food delivery company Meituan has released the weights for LongCat-2.5, one week after the model's official unveiling.

On Terminal Bench, it scores 88.4 points, landing just behind Claude Opus 5. The sample deployment runs on 194 Huawei Ascend chips, and Meituan makes no mention of Nvidia GPUs.

LongCat-2.5 is built for agents that work through long, multi-step tasks autonomously, particularly coding. According to Meituan, it comes close to top Western models on several benchmarks.

LongCat-2.5 is a mixture-of-experts model that activates only a subset of its specialized subnetworks for each token it processes. Out of 1.6 trillion total parameters, roughly 48 billion are active per token.

Through the API (the interface one piece of software uses to talk to another), the model handles up to one million tokens at once and generates responses up to 128,000 tokens long. It understands text, images, and video, responds in text, and lets users toggle reasoning on and off.