LLM Evaluation on Arena-Hard v2.0
65.8QualitySEE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SEEJudge=Gemini 3.1 Flash-Lite2026.06 | 65.8 | 61.2 | |
| Adapted RLCRJudge=Gemini 3.1 Flash-Lite2026.06 | 63.8 | 55.6 | |
| Qwen 3-4BJudge=Gemini 3.1 Flash-Lite2026.06 | 63.1 | 52.2 | |
| SEEJudge=Claude Sonnet 4.62026.06 | 48.3 | 55.1 | |
| Adapted RLCRJudge=Claude Sonnet 4.62026.06 | 47 | 49.7 | |
| Qwen 3-4BJudge=Claude Sonnet 4.62026.06 | 46.4 | 44.3 |