Video-MMMU: leaderboard
300 college-level lecture videos across 30 subjects. Measures whether watching video helps LLMs solve problems - the 'delta-knowledge' metric. Humans gain 33% from video; top models only 15%.
Metric: Overall Score. Source: videommmu.github.io. Source: videommmu.github.io. Source: llm-stats.com. Saturation forecast: Estimated already saturated. 24 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro | 87.6 | #80 |
| 2 | Gemini 3 Flash | 86.9 | #99 |
| 3 | Kimi K2.5 | 86.6 | #144 |
| 4 | GPT-5.2 | 85.9 | #110 |
| 5 | Qwen 3.7 Plus | 85.4 | #92 |
| 6 | Gemini 3.1 Flash Lite | 84.8 | #191 |
| 7 | GPT-5 | 84.6 | #98 |
| 8 | MiniMax-M3 | 84.6 | #136 |
| 9 | Qwen 3.6 27B | 84.4 | #170 |
| 10 | Qwen 3.6 Plus | 84 | #120 |
| 11 | Qwen 3.6 35B A3B | 83.7 | #226 |
| 12 | Gemini 2.5 Pro (Preview 06-05) | 83.6 | #106 |
| 13 | O3 | 83.3 | #124 |
| 14 | Seed 1.8 | 82.7 | #139 |
| 15 | Qwen 3.5 27B | 82.3 | #187 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=video-mmmu · How It Works · Data refreshed daily, snapshot 2026-10-20.