Skip to content

Open-Sora-Plan

An open-source project to reproduce Sora with Video-VQVAE and DiT

Visit official site

What Open-Sora-Plan is

The project is described as an open-source effort to reproduce OpenAI's Sora text-to-video model and to build knowledge about Video-VQVAE (VideoGPT) plus DiT. The project was started by a joint AIGC lab of Peking University and Tuzhan, and it provides training code while inviting community pull requests.

What you can do with it

  • Reproduce the Sora text-to-video model
  • Build knowledge of Video-VQVAE (VideoGPT) and DiT
  • Train a model with the shared training code
  • Take part in open-source community collaboration

Application scenarios

  • Researchers reproduce Sora for video generation research
  • Developers train their own Video-VQVAE model with the project code
  • Students learn how to use DiT for video processing

Open-source collaboration

  • The project was initiated by a joint AIGC lab of Peking University and Tuzhan
  • The project states resources are limited and asks the community to contribute
  • Training code is provided and pull requests are welcomed

Pros & cons

✓ What we like

  • Training code is shared for others to use
  • Pull requests from the community are welcome

! What to watch out for

  • The project states its resources are currently limited
  • Reproduction work depends on community contributions

FAQ

What is the goal of the project?

The material states it reproduces OpenAI's Sora text-to-video model and builds knowledge of Video-VQVAE and DiT.

Can others contribute?

The material states training code is provided and pull requests from the community are welcome.

Last reviewed: 2026-11-15

More AI video generation tools

View all →

How we review