PLawBench: leaderboard
Rubric-based benchmark for real-world legal practice: 850 questions across 13 practical legal scenarios, each scored against expert-written rubrics.
Metric: Overall (imported). Source: github.com. Source: benchmarklist.com. Saturation forecast: Around October 2027. 29 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.8 Max | 73.2 | #35 |
| 2 | GPT-5.6 Sol (Max) | 72.3 | #19 (GPT-5.6 Sol) |
| 3 | Claude Fable 5 | 70.2 | #16 |
| 4 | GPT-5.2 | 69.67 | #110 |
| 5 | Claude Opus 4.8 | 69.6 | #37 |
| 6 | GPT-5 | 67.76 | #98 |
| 7 | Claude Opus 4.5 | 66.47 | #83 |
| 8 | Gemini 3 Pro (Preview) | 66.35 | #66 |
| 9 | Claude Sonnet 4.5 | 65.88 | #142 |
| 10 | Qwen 3 Max | 64.75 | #208 |
| 11 | Gemini 2.5 Pro | 64.05 | #152 |
| 12 | Qwen 3 235B A22B 2507 Instruct | 63.08 | #300 |
| 13 | Gemini 2.5 Flash | 62.07 | #245 |
| 14 | GLM-4.6 | 60.49 | #252 |
| 15 | Qwen 3.7 Max | 58.9 | #75 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=plawbench · How It Works · Data refreshed daily, snapshot 2026-10-20.