# Identity-consistent video generation on Identity-consistent video generation dataset 500 videos

Source: [Identity-consistent video generation on Identity-consistent video generation dataset 500 videos](<https://www.sota2.com/research/sota/identity-consistent-video-generation-on-identity-consistent-video-generation-dataset-500-videos>)

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

Data updated: 2026-03-23T12:04:50.303Z

Evaluation of models generating videos that maintain subject identity, tested on a specific dataset comprising 500 videos. Metrics focus on identity similarity (ArcSim, CurSim) and visual alignment (ViCLIP-T).

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 | Identity-consistent video generation |
| Dataset (including any stated split) | Identity-consistent video generation dataset 500 videos |
| Metric | ArcSim |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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

- [ArcSim](<https://www.sota2.com/research/sota/identity-consistent-video-generation-on-identity-consistent-video-generation-dataset-500-videos.md?metric=ArcSim>): Higher is better; unit: SCALAR.

- [CurSim](<https://www.sota2.com/research/sota/identity-consistent-video-generation-on-identity-consistent-video-generation-dataset-500-videos.md?metric=CurSim>): Higher is better; unit: SCALAR.

- [ViCLIP-T](<https://www.sota2.com/research/sota/identity-consistent-video-generation-on-identity-consistent-video-generation-dataset-500-videos.md?metric=ViCLIP-T>): Higher is better; unit: SCALAR.

## Recorded results

| Method | ArcSim | Result date | Report paper |
| --- | --- | --- | --- |
| LumosX | 0.51 | Not recorded | 69c0c9150636e5bf4afc3cc2 |
| Phantom | 0.508 | Not recorded | 69c0c9150636e5bf4afc3cc2 |
| SkyReels-A2 | 0.382 | Not recorded | 69c0c9150636e5bf4afc3cc2 |

### LumosX

ArcSim: 0.51. Result ID: 69c12c5647f34ce924a9e124.

[Reporting paper](<https://www.sota2.com/research/paper/69c0c9150636e5bf4afc3cc2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69c0c9150636e5bf4afc3cc2.md>)

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

### Phantom

ArcSim: 0.508. Result ID: 69c12c5647f34ce924a9e123.

[Reporting paper](<https://www.sota2.com/research/paper/69c0c9150636e5bf4afc3cc2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77ed1e209131cea7614e.md>)

[Implementation](<https://phantom-video.github.io/Phantom/>)

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

### SkyReels-A2

ArcSim: 0.382. Result ID: 69c12c5647f34ce924a9e122.

[Reporting paper](<https://www.sota2.com/research/paper/69c0c9150636e5bf4afc3cc2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77cc1e209131cea7566f.md>)

[Implementation](<https://github.com/SkyworkAI/SkyReels-A2>)

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.
