LexRubric - Legal Accuracy: leaderboard
Metric: Dimension score rate (%): points earned from expert-written atomic rubric items of the legal accuracy dimension (weighted -10 to +10; negative items subtract) divided by the maximum positive score, averaged over instances; a Qwen3.6-27B judge checks each item, temperature 0.6, 16k output tokens; can fall below 0; higher is better (instances without a positive item in the dimension are left out). Source: arxiv.org. Saturation forecast: Around August 2027. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.6 Max Preview | 69.51 |
| 2 | Kimi K2.6 | 68.53 |
| 3 | Qwen 3 Max | 68.45 |
| 4 | Kimi K2.5 | 66.6 |
| 5 | GLM-5.1 | 62.13 |
| 6 | Claude Sonnet 4.6 | 60.6 |
| 7 | Qwen 3.5 397B A17B | 59.46 |
| 8 | GLM-5 | 56.92 |
| 9 | DeepSeek V4 Flash | 55.34 |
| 10 | DeepSeek V4 Pro | 54.21 |
| 11 | DeepSeek V3.2 | 54.17 |
| 12 | DeepSeek R1 | 52.97 |
| 13 | GPT-5.2 | 37.75 |
Interactive version: theaggregate.ai/benchmark?slug=lexrubric-legal-accuracy · How It Works · Data refreshed daily, snapshot 2026-09-29.