SWE-2 trails Fable 5.1 by one point, and by 28 on Terminal-Bench 4
Cognition shipped SWE-2 on 10 September 2026, a Kimi K3 post-trained with RL. It sits one point under Fable 5.1 on Cognition's own benchmark and at less than half of ...
Tag: coding agents
OpenAI’s 3.1 agent-workdays per human day is not a 3.1x gain
OpenAI published its research acceleration numbers on 6 September. The ratio is real, the crossover happened in June 2026, and the post itself explains why 3.1 is not a productivity ...