Benchmark
Harvey Held-Out Legal Set · Criterion pass
Scores are percentages. Higher is better.
Benchmark sourceScores from BenchBoard docs.
- Models measured
- 2
- Scores recorded
- 3
- Sources
- 1
- IQ weight
- No weight published
Score distribution
- 92.1%
- Mean
- 92.1%
- Median
- 0.776%
- Std dev
- 91.2%
- 10th percentile
- 93.1%
- 90th percentile
- 91.2%
- Min
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 5.5adaptive thinking · max effort · default sampling · 5-trial mean · context ≤1M unless noted · 120-task held-out set · max effort · mean criterion-pass· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 93.1% |
| 2 | Claude Sonnet 5.5adaptive thinking · max effort · default sampling · 5-trial mean · context ≤1M unless noted · 120-task held-out set · high effort · mean criterion-pass· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 92.1% |
| 3 | Claude Opus 5.5Opus 5.5 System Card capability section · 120-task held-out set · max · mean criterion-pass· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 91.2% |