LLM Evaluation on WildBench v2
64.9Quality ScoreSEE
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SEEJudge=Gemini 3.1 Flash-Lite2026.06 | 64.9 | 62 | |
| Adapted RLCRJudge=Gemini 3.1 Flash-Lite2026.06 | 63.7 | 56 | |
| Qwen 3-4BJudge=Gemini 3.1 Flash-Lite2026.06 | 63.3 | 52.4 | |
| SEEJudge=Claude Sonnet 4.62026.06 | 48.8 | 56.4 | |
| Adapted RLCRJudge=Claude Sonnet 4.62026.06 | 47.1 | 50 | |
| Qwen 3-4BJudge=Claude Sonnet 4.62026.06 | 46.5 | 43 |