LongCat 2.0 is a sparse mixture-of-experts language model from Meituan, with 48B active parameters out of 1.6T total. It is suited for coding, repository-level changes, long-horizon problem solving,...
Every provider for these models has hit its rate limit. They come back automatically once the limits lift.