Composer 2.5 being Pareto dominant in coding per CursorBench is important. This is after only a few weeks of supplemental training and/or RL in the Colossus 2 cluster.   The 1.5 trillion parameter version of Grok will likely be a much better base model than Kimi. We shall see.

Composer 2.5 has officially launched as the company’s most powerful model yet. The new version is designed to be more intelligent, more reliable with complex instructions and significantly better at handling long-running tasks without losing context. To celebrate the launch, included model usage is being doubled for the next week. AI models are rapidly evolving from quick-response assistants into systems capable of sustained, high-level work.