← sources

Opus 5.5 leads the code review benchmark

source: reddit (reddit:ClaudeCode)published 2026-09-25 16:30ingested 2026-09-26 02:38

retained text

So 5.5 leads this code review benchmark, intersting to see gpt-6 sol with almost similar score at less than 1/3rd the cost

document fetched 2026-09-26 00:40 · canonical https://reddit.com/r/ClaudeCode/comments/1wq0y2d/opus_55_leads_the_code_review_benchmark