Scott Wu

Stories

Cognition's SWE-2 coding model nears frontier AI performance at lower cost

As of September 2026: Leading the company behind SWE-2

Cognition released SWE-2, a coding model that scored 92.8 on Terminal-Bench 2.1 and 50.0 on FrontierCode 1.1 Main, within one point of Anthropic's Claude Fable 5.1. Cognition claims SWE-2 costs 64% less than Fable 5.1 for that performance.

Updated Sep 10