Skip to content

Open-Sora-Plan-v1.1.0

An open text-to-video model with an optimised CausalVideoVAE

Visit official site

What Open-Sora-Plan-v1.1.0 is

The model is described as a text-to-video generation model built by a team from Peking University and first released as v1.0.0 in April 2024. Version 1.1.0 brings better compressed visual representation, higher generation quality and longer video generation through an optimised CausalVideoVAE architecture.

What you can do with it

  • Use the optimised CausalVideoVAE for stronger inference efficiency
  • Train with high-quality visual data and captions
  • Apply GAN loss to keep high-frequency details and cut grid artifacts
  • Access the released code, data and models

Application scenarios

  • Researchers generate videos from descriptive text for academic demos
  • Content creators make video material for social platforms
  • Companies produce promotional videos for their products

Getting started

  • Visit the GitHub page to read about the project
  • Read the docs for access to code, data and models
  • Set up a development environment and install dependencies
  • Download and run the training scripts to generate videos

Pros & cons

✓ What we like

  • Longer video generation and better quality than v1.0.0
  • Minimal design and data efficiency kept from v1.0.0

! What to watch out for

  • Running the model asks for a development environment and dependencies
  • Performance is described as similar to the Sora base model

FAQ

When did v1.0.0 appear?

The material states it was first released in April 2024.

What does the open release contain?

Code, data and models are published for the community.

Last reviewed: 2026-11-15

More AI video generation tools

View all →

How we review