LegalEvalHub - LegalBench: leaderboard
LegalBench run across all 162 tasks under LegalEvalHub's own harness, ranked by average raw metric. Distinct from the HELM LegalBench board we also carry: different harness, different roster, and no task subsetting.
Metric: Raw Metric Average (%, 162 tasks). Source: www.legalevalhub.ai. Saturation forecast: Estimated already saturated. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.1 70B Instruct | 75.8 | #561 |
| 2 | Claude 3.5 Haiku (20241022) | 72.49 | #588 |
| 3 | GPT-4o Mini | 70.85 | #599 |
| 4 | Gemma 2 27B (IT) | 66.38 | #645 |
| 5 | Claude 3 Haiku (20240307) | 57.38 | #771 |
| 6 | GPT-4.1 Nano | 52.5 | #731 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=legalevalhub-legalbench · How It Works · Data refreshed daily, snapshot 2026-10-20.