Policy Evaluation on PolicyBench Level 2 (CN)
62.92AccuracyDeepseek R1
Evaluation Results
| Method | Links | |
|---|---|---|
| Deepseek R12026.04 | 62.92 | |
| Gemini 2.5Model Variant=Gemini-2.5-Flash2026.04 | 60.57 | |
| Claude 3.7Model Variant=Claude-3.7-Sonnet2026.04 | 60.47 | |
| QwQ 32B2026.04 | 59.79 | |
| Claude 3.5Model Variant=Claude-3.5-Sonnet2026.04 | 59.74 | |
| LLaMA 42026.04 | 56.56 | |
| Gemini 2.0Model Variant=Gemini-2.0-Flash2026.04 | 56.39 | |
| GPT-4o2026.04 | 56.34 | |
| o4-mini2026.04 | 55.81 | |
| Gemma 3-27B2026.04 | 55.56 | |
| Deepseek V32026.04 | 55.51 |