Meituan Releases LongCat-2.5 Weights Under Full MIT License
Meituan released the full weights of LongCat-2.5, a 1.6-trillion-parameter MoE model with a one-million-token context, under an unrestricted MIT license on Hugging Face.
Meituan released the weights of its LongCat-2.5 model on Hugging Face on November 9 under a full MIT license. The company describes it as a 1.6-trillion-parameter mixture-of-experts model with roughly 48 billion active parameters, natively multimodal, with a context window of one million tokens. The model was announced on November 5 with weights marked as "coming soon"; the November 9 release is the next stage of that rollout.
An MIT license carries no commercial restrictions, meaning a frontier-scale model can now be downloaded, modified, and used commercially without restriction, a reinforcement of the open-weight trend among large model developers. Because the model is exceptionally large, running it is realistic only for operators with GPU capacity at significant scale rather than small infrastructure. However, the availability of weights opens the possibility of community-driven distillation and quantization efforts.