Skip to content

Benchmark

VBench-Long object class

Scores are percentages.

Models measured
9
Scores recorded
9
Sources
1
IQ weight
0.51

Score distribution

89.8%
Mean
88.4%
Median
4.7%
Std dev
83.5%
10th percentile
93.9%
90th percentile
83.5%
Min
83.5%97.4%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.51 for video-generation models, measured over 8 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Vidu Q12025-04-17· VBench data docs97.4%
2Sora· VBench data docs93.9%
3Veo 3· VBench data docs93.9%
4Kling V1.6· VBench data docs93.3%
5Vidu Q4· VBench data docs88.4%
6Video 01· VBench data docs87.8%
7Mochi 1· VBench data docs86.5%
8Hunyuan Video2025-05-22· VBench data docs83.5%
9LTX Video5s 768×512· VBench data docs83.5%

Data credits: VBench (Vchitect), MIT (licence)