MME-CoT - CoT Quality F1: leaderboard
Each model answers MME-CoT's visual questions step by step; on its 837 reasoning questions a GPT-4o judge matches the response against annotated key steps (recall) and checks each image observation and inference it states (precision), and the score is their F1. Unlike answer-accuracy boards such as MathVista, it grades the reasoning chain, not the final answer.
Metric: F1 (%). Source: arxiv.org. Saturation forecast: Around February 2027. 7 models tracked. Updated 17 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | QVQ-72B-Preview | 62 | #1173 |
| 2 | Virgo-72B | 60.8 | #1354 |
| 3 | InternVL2.5-78B-MPO | 52.7 | #1203 |
| 4 | InternVL2.5-8B-MPO | 43 | #2026 |
| 5 | InternVL2.5-8B | 41.1 | #2401 |
| 6 | MiniCPM-V-2.6 | 39.8 | #3523 |
| 7 | LLaVA-CoT | 34.9 | #3426 |
No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.
Interactive version: theaggregate.ai/benchmark?slug=mme-cot-cot-quality-f1 · How It Works · Data refreshed daily, snapshot 2026-11-17.