Z.aiCogVideoX
CogVideoX-3
A hosted video model with text, image and endpoint-frame inputs.
Practical use
Practical guidance. Choose a simple action that can fit between the supplied frames. Review the middle of the clip as closely as the endpoints, since a convincing start and finish can hide a broken transition.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | API |
| Model identifiers | cogvideox-3 |
| Input | Text, an image, or starting and ending frames.CogVideoX-3 official documentation |
| Output | Video through an asynchronous generation API.CogVideoX-3 official documentation |
Related models
- CogVideoX1.5-5B
Z.ai ยท CogVideoX
A five-billion-parameter video model with text-to-video and image-to-video checkpoints.
Launch date not confirmedRestricted accessDownloadable