meituan-longcat/LongCat-2.0

Hugging Face Models Trending Models

Summary

LongCat-2.0 is a large-scale MoE language model with 1.6 trillion total parameters and ~48B activated per token, trained on AI ASIC superpods with 1M-context data. It achieves strong performance on coding and agentic tasks.

Tags: region:us
Original Article
View Cached Full Text

Cached at: 06/30/26, 11:30 PM

meituan-longcat/LongCat-2.0 · Hugging Face

Source: https://huggingface.co/meituan-longcat/LongCat-2.0 LongCat-2.0


Hugging Face

WechatTwitter Follow

License

Tech Blog📄

https://huggingface.co/meituan-longcat/LongCat-2.0#model-introductionModel Introduction

We introduce LongCat-2.0, a large-scale MoE language model with1.6 trillion total parametersand ~48 billion activated per token — a substantial step up from previous LongCat models, accompanied by several architectural improvements.

Both the full training run and the large-scale deployment are built entirely onAI ASIC superpods. Pretraining spans millions of accelerator-hours across more than 35 trillion tokens, with no rollbacks or irrecoverable loss spikes — demonstrating that we have the capability to conduct frontier-scale training on alternative hardware platforms.

To strengthen the model on long-horizon tasks, we introduce LongCat Sparse Attention and train LongCat-2.0 on hundreds of billions of tokens of1M-contextdata. Together with dedicated post-training, this gives LongCat-2.0 strong performance on coding and agentic tasks.


🏋️Model weights coming soon— stay tuned!

Similar Articles

LongCat-2.0

Product Hunt

LongCat-2.0 is a 1.6 trillion parameter mixture-of-experts model trained entirely on custom AI ASICs.