long context benchmark
MLVU
A comprehensive benchmark for multi-task long video understanding that evaluates multimodal large language models on videos ranging from 3 minutes to 2 hours across 9 distinct tasks including reasoning, captioning, recognition, and summarization.
Updated Aug 11, 2026
Models10
Model coverage10
MetricScore
EvidenceB
MLVU Ranking
Higher score ranks better on this benchmark.
| 01 | AC | 87.4% | 100.0% | 10 | C | |
| 02 | AC | 87.3% | 88.9% | 10 | C | |
| 03 | AC | 86.7% | 77.8% | 10 | C | |
| 04 | AC | 86.6% | 66.7% | 10 | C | |
| 05 | AC | 86.2% | 55.6% | 10 | C | |
| 06 | AC | 85.9% | 44.4% | 10 | C | |
| 07 | AC | 85.6% | 33.3% | 10 | C | |
| 08 | AC | 84.3% | 22.2% | 10 | C | |
| 09 | AC | 83.8% | 11.1% | 10 | C | |
| 10 | AC | 70.2% | 0.0% | 10 | C |
MLVU Score Distribution
A closer view of the leading scores on this benchmark.
MLVU Highlights
The leading models and scores on this benchmark.
What is MLVU?
What MLVU measures and how its scores work.
A comprehensive benchmark for multi-task long video understanding that evaluates multimodal large language models on videos ranging from 3 minutes to 2 hours across 9 distinct tasks including reasoning, captioning, recognition, and summarization.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- MLVU
- Modality
- multimodal
- Primary category
- long context
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- mlvu|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about MLVU.
Which model scores highest on MLVU?
Qwen3.7-Plus is currently ranked first with 87.4%.
What does MLVU measure?
A comprehensive benchmark for multi-task long video understanding that evaluates multimodal large language models on videos ranging from 3 minutes to 2 hours across 9 distinct tasks including reasoning, captioning, recognition, and summarization.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
10 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.