RMGAP - Listwise LLM Judge: leaderboard
Metric: Pairwise accuracy (%; chosen response scored above a rejected one, prompts asking for one of four styles, mean of chat, writing, reasoning and safety; the model ranks the four shuffled responses in one call, ranks mapped to scores 4-1). Source: arxiv.org. Saturation forecast: Around December 2026. 5 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 235B A22B 2507 | 69.77 |
| 2 | DeepSeek V3.2 | 67.83 |
| 3 | Qwen 2.5 7B Instruct | 58.23 |
| 4 | Qwen 3 8B | 57.84 |
| 5 | Llama 3.1 8B Instruct | 50.08 |
Interactive version: theaggregate.ai/benchmark?slug=rmgap-listwise-llm-judge · How It Works · Data refreshed daily, snapshot 2026-09-26.