AI Change Tracker
4

Claude Opus 4.8 posts 78.2 on SWE-bench Verified

2026-08-20

Anthropic published a 78.2% SWE-bench Verified score for Opus 4.8 with a minimal scaffold, a 1.4-point jump over the previous top entry. Lab-reported; the leaderboard maintainers have queued an independent run.

Significance 4: New top score on the primary coding axis source; lab-reported cap applies.

Benchmark result

Sources

Models

Affected axes

Coding

← Back to feed