LexRubric - Reasoning and Logic: leaderboard
LexRubric (2026) scores open-ended answers to 649 Chinese legal questions from everyday legal consultation and judicial examination against expert-written atomic rubric items, each checked by an LLM judge. This board counts only the rubric items of the reasoning and logic dimension. The score rate is earned rubric points (negative items subtract) over the maximum positive score in percent, averaged over instances; higher is better.
Metric: Dimension score rate (%): points earned from expert-written atomic rubric items of the reasoning and logic 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 July 2027. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Kimi K2.6 | 79.08 | #178 |
| 2 | Qwen 3.6 Max Preview | 77.14 | #159 |
| 3 | Qwen 3 Max | 77.07 | #403 |
| 4 | Kimi K2.5 | 76.5 | #256 |
| 5 | GLM-5.1 | 74.72 | #179 |
| 6 | Claude Sonnet 4.6 | 72.24 | #161 |
| 7 | Qwen 3.5 397B A17B | 72.17 | #246 |
| 8 | GLM-5 | 68.24 | #260 |
| 9 | DeepSeek V4 Pro | 65.86 | #188 |
| 10 | DeepSeek V4 Flash | 65.71 | #258 |
| 11 | GPT-5.2 | 63.87 | #176 |
| 12 | DeepSeek R1 | 62.43 | #473 |
| 13 | DeepSeek V3.2 | 61.36 | #406 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=lexrubric-reasoning-and-logic · How It Works · Data refreshed daily, snapshot 2026-11-13.