Meituan releases LongCat-2.5 agent model trained on Chinese hardware
Chinese company Meituan has released the weights for LongCat-2.5, a 1.6-trillion-parameter model designed to handle complex coding and agent tasks autonomously. The model was trained entirely on domestic hardware including Huawei Ascend chips, achieving benchmark scores close to Claude Opus 5 without Nvidia processors. It supports multimodal inputs across text, images, and video, handling a context window of up to one million tokens.
Original
Meituan releases a 1.6-trillion-parameter agent model that never touched an Nvidia GPU
Meituan has released the weights for LongCat-2.5, a 1.6-trillion-parameter model built for AI agents and trained entirely on Chinese hardware. 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. The article Meituan releases a 1.6-trillion-parameter agent model that never…