CWE-bench v2: More coverage, quality, and transparency
Comments
Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules
Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules
Summary
CWE-bench v2 from Collinear AI releases a public set of 100 tasks testing whether AI agents can find and fix security weaknesses in real open-source repositories without breaking existing functionality. Scoring is fully programmatic and reproducible, and the tasks span 91 CWEs.