Coding benchmark
Bug Hunt Bench leaderboard
Every model the catalog carries a published Bug Hunt Bench value for, ranked by that value.
CategoryCoding
MeasureStrict planted bugs fixed
Tasks105 planted bugs across two production TypeScript repositories
DifficultyBlind production-repository bug finding and repair
Published byBug Hunt Bench method, caveats, and definitions
A blind-graded coding-agent benchmark with 105 planted bugs across two production TypeScript repositories.
Bug Hunt Bench ranking
4 models with a published Bug Hunt Bench value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Strict planted bugs fixed |
|---|---|---|---|
| 1 |  Claude Fable 5.1 | Anthropic | 43 |
| 2 |  GPT-5.6 Sol | OpenAI | 42 |
| 3 | Claude Opus 5 | Anthropic | 27 |
| 3 |  Grok 4.6 | xAI | 27 |
Evidence key: Observed
Rows are ordered by the value Bug Hunt Bench method, caveats, and definitions published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.