Context Learning in Video Generation on CLVG-Bench
52.64P.R. ScoreSeedance 2.0
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Seedance 2.0Model Type=Proprietary, Supported Inputs=video, audio, and image2026.04 | 52.64 | 61.25 | 67.65 | 63.54 | 0.4337 | 21.25 | |
| Sora 2Model Type=Proprietary, Supported Inputs=video and image2026.04 | 49.45 | 46.38 | 52.49 | 48.91 | 0.4586 | 38.96 | |
| Wan 2.2 14BModel Type=Open Source, Supported Inputs=image only2026.04 | 46.67 | — | — | 39 | 0.16 | 11.25 | |
| Veo 3.1 FastModel Type=Proprietary, Supported Inputs=image only2026.04 | 46.38 | — | — | 46.88 | 0.36 | 16.25 | |
| Seedance 1.5Model Type=Proprietary, Supported Inputs=image only2026.04 | 35.47 | — | — | 37.5 | 0.4 | 8.75 | |
| HunyuanVideo 13BModel Type=Open Source, Supported Inputs=image only2026.04 | 26.67 | — | — | 26 | 0.14 | 5 | |
| LTX-2Model Type=Open Source, Supported Inputs=video and image2026.04 | 11.29 | 0.72 | 2.04 | 16 | 0.02 | 2.5 | |
| UniVideoModel Type=Open Source, Supported Inputs=video and image2026.04 | 8.06 | 8.1 | 4.08 | 10 | 0.03 | 3.75 |