RMGAP - Pointwise LLM Judge - Best-of-4: leaderboard
Metric: Best-of-4 accuracy (%; chosen response scored above all three alternatives, mean of chat, writing, reasoning and safety; the model rates each response 1-10, 5 when no score can be parsed, ties count as errors). Source: arxiv.org. Saturation forecast: Around October 2027. 5 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 235B A22B 2507 | 34.61 |
| 2 | DeepSeek V3.2 | 29.15 |
| 3 | Qwen 3 8B | 21.6 |
| 4 | Qwen 2.5 7B Instruct | 20.02 |
| 5 | Llama 3.1 8B Instruct | 9.59 |
Interactive version: theaggregate.ai/benchmark?slug=rmgap-pointwise-llm-judge-best-of-4 · How It Works · Data refreshed daily, snapshot 2026-09-26.