LexGenius: leaderboard
Expert-level benchmark for legal general intelligence, covering Chinese legal understanding and reasoning across multiple task dimensions.
Metric: CoT average (imported). Source: qwenqking.github.io. Source: benchmarklist.com. Saturation forecast: Estimated already saturated. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek R1 | 64.59 | #251 |
| 2 | DeepSeek V3 | 64.39 | #321 |
| 3 | Qwen 2.5 7B | 59.38 | #922 |
| 4 | Qwen 3 8B | 55.3 | #674 |
| 5 | GPT-4.1 Nano | 53.43 | #731 |
| 6 | Qwen 3 4B | 53.19 | #829 |
| 7 | GPT-4o Mini | 52.61 | #599 |
| 8 | Qwen 2.5 1.5B | 48.22 | #1431 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=lexgenius · How It Works · Data refreshed daily, snapshot 2026-10-20.