MCJudgeBench — leaderboard
Metric: CJAR (self-reported). Source: benchmarklist.com. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 85.8 |
| 2 | Gemini 2.5 Flash Lite | 85.5 |
| 3 | Qwen 3.5 4B | 85.3 |
| 4 | Claude Haiku 4.5 | 84.8 |
| 5 | Claude Sonnet 4.6 | 82.8 |
| 6 | GPT-5.2 | 80.9 |
| 7 | Llama 3.2 3B Instruct | 77 |
Interactive version: theaggregate.ai/benchmark?slug=mcjudgebench · How the rankings work · Data refreshed daily, snapshot 2026-07-22.