Policy Evaluation on PolicyBench Level 1 (CN)
62.02AccuracyDeepseek R1
Evaluation Results
| Method | Links | |
|---|---|---|
| Deepseek R12026.04 | 62.02 | |
| QwQ 32B2026.04 | 55.87 | |
| Claude 3.7Model Variant=Claude-3.7-Sonnet2026.04 | 55.29 | |
| Gemini 2.5Model Variant=Gemini-2.5-Flash2026.04 | 54.06 | |
| Claude 3.5Model Variant=Claude-3.5-Sonnet2026.04 | 53.77 | |
| LLaMA 42026.04 | 49.81 | |
| Deepseek V32026.04 | 48.61 | |
| Gemini 2.0Model Variant=Gemini-2.0-Flash2026.04 | 47.87 | |
| GPT-4o2026.04 | 46.01 | |
| o4-mini2026.04 | 45.93 | |
| Gemma 3-27B2026.04 | 41.75 |