Claude Opus 4 Sets New SWE-bench Record at 72.5%
· Anthropic Announcements
Anthropic's Claude Opus 4 achieves 72.5% on SWE-bench, setting a new state-of-the-art record for autonomous software engineering capabilities....
Anthropic's Claude Opus 4 achieves 72.5% on SWE-bench, setting a new state-of-the-art record for autonomous software engineering capabilities.