Cognition(@cognition):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 scale

Cognition

@cognition

Makers of Devin, the first AI software engineer. We are an applied AI lab building end-to-end software agents. Join us:

Joined January 2024

6 Following 171K Followers

Cognition@cognition

2026.07.08 15:47

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 scale