Skip to content

Benchmark

VBench 1.0 Semantic Score

Scores are percentages.

Models measured
9
Scores recorded
9
Sources
1
IQ weight
0.74

Score distribution

77.4%
Mean
77%
Median
5.3%
Std dev
70.8%
10th percentile
82.5%
90th percentile
70.1%
Min
70.1%87.9%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.74 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 docs87.9%
2Veo 3· VBench data docs82.5%
3Sora· VBench data docs79.4%
4Video 01· VBench data docs77.7%
5Kling V1.6· VBench data docs77%
6Hunyuan Video2025-05-22· VBench data docs76.9%
7Vidu Q4· VBench data docs74%
8LTX Video5s 768×512· VBench data docs70.8%
9Mochi 1· VBench data docs70.1%

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