Legal Responsibility Mode Classification on C-TRAIL (test)
86.4AccuracyLegal Multi-Agent Reasoning framework
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Legal Multi-Agent Reasoning frameworkCategory=Agent Methods2026.03 | 86.4 | 79.1 | 73.2 | |
| AgentsCourtCategory=Agent Methods2026.03 | 85.1 | 77.3 | 71.2 | |
| Qwen2.5-LawCategory=Legal LLMs2026.03 | 84.7 | 76.7 | 69.7 | |
| ChatLaw2-MoECategory=Legal LLMs2026.03 | 84.1 | 75.3 | 68.5 | |
| GPT-4Category=General LLMs2026.03 | 82.9 | 74.7 | 67.5 | |
| LaWGPTCategory=Legal LLMs2026.03 | 81.8 | 72.4 | 64.1 | |
| DeepSeek-R1Category=General LLMs2026.03 | 81.3 | 73.2 | 63.8 | |
| GPT-3.5Category=General LLMs2026.03 | 79.6 | 68.7 | 61.3 | |
| ReActCategory=Agent Methods2026.03 | 78.4 | 67.5 | 59.5 |