Benchmark
RevBench
Scores are percentages.
Benchmark sourceScores from Harbor Hub Leaderboards.
- Models measured
- 3
- Scores recorded
- 3
- Sources
- 1
- IQ weight
- No weight published
Score distribution
- 37.9%
- Mean
- 36.4%
- Median
- 9.3%
- Std dev
- 27.3%
- 10th percentile
- 50%
- 90th percentile
- 27.3%
- Min
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | GPT 5.5revbench· Harbor Hub Leaderboards | 50% |
| 2 | Qwen3.8 Maxrevbench· Harbor Hub Leaderboards | 36.4% |
| 3 | DeepSeek V4 Flash 0731revbench· Harbor Hub Leaderboards | 27.3% |