Model Family
Sora
OpenAI's text-to-video diffusion-transformer model, up to 60-second cinematic clips.
Definition
Sora generates high-fidelity video from text or images. It uses a diffusion transformer that treats video as patches in time. Released publicly in late 2024 inside ChatGPT, it pushed open expectations of generative video quality and length.
Common use cases
- Video content
- Filmmaking
- Prototyping