MMBench-Video
multimodal
+
+
+
+
About
MMBench-Video extends the MMBench evaluation framework to video understanding tasks, assessing vision-language models' ability to comprehend temporal visual content. It features carefully designed multiple-choice questions that evaluate video-based reasoning, temporal understanding, and multimodal comprehension across various video genres and scenarios in both English and Chinese contexts.
+
+
+
+
Evaluation Stats
Total Models3
Organizations1
Verified Results0
Self-Reported3
+
+
+
+
Benchmark Details
Max Score1
Language
en
+
+
+
+
Performance Overview
Score distribution and top performers
Score Distribution
3 models
Top Score
2.0%
Average Score
1.9%
High Performers (80%+)
0Top Organizations
#1Alibaba Cloud / Qwen Team
3 models
1.9%
+
+
+
+
Leaderboard
3 models ranked by performance on MMBench-Video
| License | Links | ||||
|---|---|---|---|---|---|
| Jan 26, 2025 | tongyi-qianwen | 2.0% | |||
| Feb 28, 2025 | Apache 2.0 | 1.9% | |||
| Jan 26, 2025 | Apache 2.0 | 1.8% |
