Muffin Study - Multi-Agent Debate - MMLU-Pro: leaderboard
Muffin study:fixed Multi-Agent Debate inference scaffold;four backbone models,mean of three independent runs;five reasoning steps in every method and three agents in voting/debate/Muffin;100Chess,100MMLU-Pro,198GPQA-Diamond,100eachOlymMATHEnglishEasy/Hard. Shared-memory async orchestration is a configuration,not another model. OlymMATH easy/hard English subsets separate;math rendering duplicates printed mean andSD;reported means selected once. Accuracy (0–1).
Metric: Accuracy (0–1). Source: arxiv.org. Saturation forecast: Around January 2027. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemma 4 31B (IT) | 0.85 | #179 |
| 2 | Qwen 3.8 27B | 0.82 | #111 |
| 3 | gemma-4-E4B-it | 0.77 | #572 |
| 4 | Qwen 3.5 4B | 0.74 | #493 |
No result here: #4 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=muffin-study-multi-agent-debate-mmlu-pro · How It Works · Data refreshed daily, snapshot 2026-10-24.