# Video Generation on VBench Custom (Perceptual Metrics)

Source: [Video Generation on VBench Custom (Perceptual Metrics)](<https://www.sota2.com/research/sota/video-generation-on-vbench-custom-perceptual-metrics>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T06:04:52.393Z

Evaluation of Video Generation models using the VBench Custom dataset, assessed using a suite of perceptual and quality metrics including Subject Consistency, Background Consistency, Motion Smoothness, and overall Quality Score.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Video Generation |
| Dataset (including any stated split) | VBench Custom |
| Metric | Subject Consistency |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 11 of 11 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Subject Consistency](<https://www.sota2.com/research/sota/video-generation-on-vbench-custom-perceptual-metrics.md?metric=Subject%20Consistency>): Higher is better; unit: PERCENT.

- [Background Consistency](<https://www.sota2.com/research/sota/video-generation-on-vbench-custom-perceptual-metrics.md?metric=Background%20Consistency>): Higher is better; unit: PERCENT.

- [Motion Smoothness](<https://www.sota2.com/research/sota/video-generation-on-vbench-custom-perceptual-metrics.md?metric=Motion%20Smoothness>): Higher is better; unit: PERCENT.

- [Aesthetic Quality](<https://www.sota2.com/research/sota/video-generation-on-vbench-custom-perceptual-metrics.md?metric=Aesthetic%20Quality>): Higher is better; unit: PERCENT.

- [Imaging Quality](<https://www.sota2.com/research/sota/video-generation-on-vbench-custom-perceptual-metrics.md?metric=Imaging%20Quality>): Higher is better; unit: PERCENT.

- [Dynamic Degree](<https://www.sota2.com/research/sota/video-generation-on-vbench-custom-perceptual-metrics.md?metric=Dynamic%20Degree>): Higher is better; unit: PERCENT.

- [Quality Score](<https://www.sota2.com/research/sota/video-generation-on-vbench-custom-perceptual-metrics.md?metric=Quality%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Subject Consistency | Result date | Report paper |
| --- | --- | --- | --- |
| Runway Gen-2 | 97.61 | Not recorded | 68fa78c41e209131cea7a639 |
| Pika | 96.76 | Not recorded | 68fa78c41e209131cea7a639 |
| Show-1 | 95.53 | Not recorded | 68fa78c41e209131cea7a639 |
| VideoCrafter-1.0 | 95.1 | Not recorded | 68fa78c41e209131cea7a639 |
| Sora | 94.96 | Not recorded | 68fa78c41e209131cea7a639 |
| LinGen | 94 | Not recorded | 68fa78c41e209131cea7a639 |
| CogVideo | 92.19 | Not recorded | 68fa78c41e209131cea7a639 |
| LaVie-Interpolation | 92 | Not recorded | 68fa78c41e209131cea7a639 |
| LaVie | 91.41 | Not recorded | 68fa78c41e209131cea7a639 |
| ModelScope | 89.87 | Not recorded | 68fa78c41e209131cea7a639 |
| VideoCrafter-0.9 | 86.24 | Not recorded | 68fa78c41e209131cea7a639 |

### Runway Gen-2

Subject Consistency: 97.61. Result ID: 6995a1943badfaffe69e5708.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a639.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Pika

Subject Consistency: 96.76. Result ID: 6995a1943badfaffe69e5709.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a639.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Show-1

Subject Consistency: 95.53. Result ID: 6995a1943badfaffe69e570b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a639.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78f11e209131cea7b5a6.md>)

[Implementation](<https://github.com/showlab/Show-1>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### VideoCrafter-1.0

Subject Consistency: 95.1. Result ID: 6995a1943badfaffe69e570a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a639.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71061e209131cea4ea6d.md>)

[Implementation](<https://github.com/AILab-CVC/VideoCrafter>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Sora

Subject Consistency: 94.96. Result ID: 6995a1943badfaffe69e5707.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a639.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### LinGen

Subject Consistency: 94. Result ID: 6995a1943badfaffe69e5711.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a639.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a639.md>)

[Implementation](<https://lineargen.github.io/>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### CogVideo

Subject Consistency: 92.19. Result ID: 6995a1943badfaffe69e5710.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a639.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ca31e209131cea351bf.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### LaVie-Interpolation

Subject Consistency: 92. Result ID: 6995a1943badfaffe69e570c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a639.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70ae1e209131cea4c7e4.md>)

[Implementation](<https://vchitect.github.io/LaVie-project/>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### LaVie

Subject Consistency: 91.41. Result ID: 6995a1943badfaffe69e570d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a639.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70ae1e209131cea4c7e4.md>)

[Implementation](<https://vchitect.github.io/LaVie-project/>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### ModelScope

Subject Consistency: 89.87. Result ID: 6995a1943badfaffe69e570e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a639.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70451e209131cea4a0d9.md>)

[Implementation](<https://github.com/modelscope/modelscope/blob/master/modelscope/models/multi_modal/video_synthesis>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### VideoCrafter-0.9

Subject Consistency: 86.24. Result ID: 6995a1943badfaffe69e570f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a639.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71061e209131cea4ea6d.md>)

[Implementation](<https://github.com/AILab-CVC/VideoCrafter>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
