X · @teortaxesTex
· X / Twitter
China is still behind in post-training
China is still behind in post-trainingCognition: Introducing SWE-1.7, the most capable model we’ve trained yet.It scores within a few points of the strongest frontier models at a fraction of the cost, and is now available at 1000 tok/s.RL is not hitting its limit: after refining our recipe, we keep seeing gains as we s