Skip to content

Benchmark

VBench 1.0 overall consistency

Scores are percentages.

Models measured
9
Scores recorded
9
Sources
1
IQ weight
0.48

Score distribution

26.5%
Mean
26.5%
Median
0.87%
Std dev
25.2%
10th percentile
27.2%
90th percentile
25.2%
Min
25.2%27.9%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.48 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
1Veo 3· VBench data docs27.9%
2Vidu Q12025-04-17· VBench data docs27.2%
3Video 01· VBench data docs27.1%
4Hunyuan Video2025-05-22· VBench data docs27%
5Vidu Q4· VBench data docs26.5%
6Sora· VBench data docs26.3%
7Kling V1.6· VBench data docs26%
8LTX Video5s 768×512· VBench data docs25.2%
9Mochi 1· VBench data docs25.2%

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