Benchmark
Bug Hunt Bench
A benchmark of coding models finding and fixing hidden bugs in production codebases.
Published leaderboard
Explore scores and evaluation settings on the publisher’s leaderboard.
Open leaderboardCoverage
The current repository reports 105 planted bugs in two codebases, with results for 47 models updated October 19, 2026.
Methodology
Models work through agentic coding tools; diffs are blind-graded against withheld answer keys. Report model, harness, effort, run count, and benchmark version together.