multimodal benchmark
MVBench
A comprehensive multi-modal video understanding benchmark covering 20 challenging video tasks that require temporal understanding beyond single-frame analysis. Tasks span from perception to cognition, including action recognition, temporal reasoning, spatial reasoning, object interaction, scene transition, and counterfactual inference. Uses a novel static-to-dynamic method to systematically generate video tasks from existing annotations.
Updated Aug 11, 2026
Models17
Model coverage17
MetricScore
EvidenceB
MVBench Ranking
Higher score ranks better on this benchmark.
MVBench Score Distribution
A closer view of the leading scores on this benchmark.
MVBench Highlights
The leading models and scores on this benchmark.
What is MVBench?
What MVBench measures and how its scores work.
A comprehensive multi-modal video understanding benchmark covering 20 challenging video tasks that require temporal understanding beyond single-frame analysis. Tasks span from perception to cognition, including action recognition, temporal reasoning, spatial reasoning, object interaction, scene transition, and counterfactual inference. Uses a novel static-to-dynamic method to systematically generate video tasks from existing annotations.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- MVBench
- Modality
- multimodal
- Primary category
- multimodal
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- mvbench|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about MVBench.
Which model scores highest on MVBench?
Qwen3.5-122B-A10B is currently ranked first with 76.6%.
What does MVBench measure?
A comprehensive multi-modal video understanding benchmark covering 20 challenging video tasks that require temporal understanding beyond single-frame analysis. Tasks span from perception to cognition, including action recognition, temporal reasoning, spatial reasoning, object interaction, scene transition, and counterfactual inference. Uses a novel static-to-dynamic method to systematically generate video tasks from existing annotations.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
17 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.