Multimodal Multi-disciplinary Video Understanding

A benchmark for evaluating multimodal models on video understanding tasks across multiple disciplines, emphasizing temporal reasoning and comprehension over video content.

7 models

Higher is better.

1Qwen3.8 MaxAlibaba82.4
2GLM-5.3-FlashZ.AI80.5
3Kimi K2.5Moonshot AI80.4
4dots3-note PreviewDots Studio79.9
5Qwen3.5-122B-A10BAlibaba74.7
6Qwen3.5-27BAlibaba73.3
7Qwen3.5-35B-A3BAlibaba72.3

Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric: Video reasoning benchmark
Tasks: Video understanding
Authors: MMVU benchmark maintainers (2026)
Source: Kimi K2.5 benchmark release surface ↗

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.