SGLang(@sgl_project):Congrats to the @cognition team on the SWE-2 launch! It’s especially exciting to see SpecForge helping improve speculative decoding acceptance within the training stack. We’re proud to have contributed to this effort 🙌

2026.09.11 17:23

Congrats to the @cognition team on the SWE-2 launch! It’s especially exciting to see SpecForge helping improve speculative decoding acceptance within the training stack. We’re proud to have contributed to this effort 🙌

Cognition@cognition

2026.09.10 15:21

Introducing SWE-2, our closest model yet to the frontier. On leading evals, it scores on par with recent frontier models – at up to 70% lower cost. We scaled RL to multiple trillions of parameters, with a refined recipe that pushes the Pareto curve on both capabilities & cost.

커뮤니티로 전달