Meituan releases LongCat-2.5, a 1.6-trillion-parameter multimodal model
Meituan released LongCat-2.5, a 1.6-trillion-parameter multimodal model designed for long-horizon agentic tasks and visual coding, available through its chat service and API platform.
By GennnAi··1 min read
According to press coverage, the mixture-of-experts model activates about 48 billion parameters per token. Open releases on GitHub and Hugging Face are planned. Meituan presents results competitive with some Claude Opus variants, but independent reporting says other models still lead several of its own comparisons. The release follows a preview published in September 2026. The model targets complex multi-step jobs in which the AI acts autonomously rather than answering single questions.
Source: Meituan LongCat, published on November 2, 2026.