MCJudgeBench - Macro-F1: leaderboard
Metric: Macro-F1 (%, times 100) over the yes, partial and no classes of constraint-level verdicts: the judge sees an instruction, a candidate response and its explicit constraint list and labels each constraint yes, partial or no, over the 653 constraints of 141 multi-constraint instructions of MCJudgeBench; deterministic decoding at temperature 0; higher is better. Source: arxiv.org. Saturation forecast: Around 2029. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.6 (Thinking) | 63.7 |
| 2 | GPT-5.2 | 61.6 |
| 3 | Claude Sonnet 4.6 | 60.7 |
| 4 | Claude Haiku 4.5 (Thinking) | 60.3 |
| 5 | Gemini 3.1 Pro (Preview) | 59.8 |
| 6 | GPT-5.2 (Non-reasoning) | 59.2 |
| 7 | Claude Haiku 4.5 | 58.5 |
| 8 | Gemini 2.5 Flash Lite (Thinking) | 56.9 |
| 9 | Qwen 3.5 4B (Non-reasoning) | 52.9 |
| 10 | Gemini 2.5 Flash Lite | 51.8 |
| 11 | Llama 3.2 3B Instruct | 44.1 |
Interactive version: theaggregate.ai/benchmark?slug=mcjudgebench-macro-f1 · How It Works · Data refreshed daily, snapshot 2026-10-07.