Benchmark
Vals CaseLaw v2
Scores are percentages.
Benchmark sourceScores from Vals AI.
- Models measured
- 53
- Scores recorded
- 53
- Sources
- 1
- IQ weight
- 0.76
Score distribution
- 59%
- Mean
- 59.4%
- Median
- 7.1%
- Std dev
- 51.6%
- 10th percentile
- 66.5%
- 90th percentile
- 43.8%
- Min
This benchmark has a weight of 0.76 for chat models, measured over 53 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 51 | Qwen3.6 Max Preview· Vals AI | 47.9% |
| 52 | Mistral Medium 3.5· Vals AI | 44.2% |
| 53 | GPT OSS 20b· Vals AI | 43.8% |
Showing 51–53 of 53 results.