MVPBench - Professional Action Evaluation: leaderboard
Metric: MVPBench normalized proficiency score (accuracy minus chance, divided by one minus chance, times 100, so 0 is chance level and 100 is perfect) on its 70 multi-video questions choosing the better-executed of two AQA-7 sports clips; chance is 1/2; answers extracted by rules and GPT-4-turbo; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 15 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 37.14 | #333 |
| 2 | Gemini 2.5 Flash | 28.74 | #237 |
| 3 | InternVL3-78B | 14.28 | #345 |
| 4 | Qwen 2.5 VL 72B Instruct | 14.28 | #364 |
| 5 | InternVL3-8B | 5.72 | #606 |
| 6 | InternVL3-38B | 5.72 | #395 |
| 7 | Qwen 2.5 VL 7B Instruct | 5.72 | #643 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=mvpbench-professional-action-evaluation · How It Works · Data refreshed daily, snapshot 2026-10-11.