MVBench: leaderboard
MVBench: Evaluates temporal, video, speech, or audio understanding beyond static text and image inputs.
Metric: Average accuracy (%). Source: github.com. Source: benchmarklist.com. Saturation forecast: Around January 2027. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.7 Flash | 82.2 | #23 |
| 2 | GLM-5.3 Flash | 77.8 | #65 |
| 3 | GPT-5.6 Terra | 75 | #41 |
| 4 | DeepSeek-V4-Flash-Vision-Exp | 69.4 | #76 |
| 5 | Claude Opus 4.8 | 67.1 | #37 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mvbench · How It Works · Data refreshed daily, snapshot 2026-10-20.