List your company

Autonomous Software Engineering on SWE-Bench Pro (test)

32.6Resolution RateSonnet
29.89630.59831.332.002May 4, 2026

Evaluation results

MethodLinks
SonnetSubagent Config.=Sonnet, Main Agent Model=Claude Opus 4.62026.0532.6756197740.66100
SFT-4BSubagent Config.=SFT-4B, Main Agent Model=Claude Opus 4.62026.0532.4727177271.11798.5
OpusSubagent Config.=Opus, Main Agent Model=Claude Opus 4.62026.0532729197480.54100
Terminus-4BSubagent Config.=Terminus-4B, Main Agent Model=Claude Opus 4.62026.0531.57301873011497.9
Vanilla-4BSubagent Config.=Vanilla-4B, Main Agent Model=Claude Opus 4.62026.0530.484068401.92799
No SubagentSubagent Config.=No Subagent, Main Agent Model=Claude Opus 4.62026.0530836—8363.8——