🎬 LongLive: Real-time Interactive Long Video Generation

💡 TLDR: Turn interactive prompts into long videos—instantly, as you type!

LongLive: Real-time Interactive Long Video Generation [Paper]
Shuai Yang, Wei Huang, Ruihang Chu, Yicheng Xiao, Yuyang Zhao, Xianbang Wang, Muyang Li, Enze Xie, Yingcong Chen, Yao Lu, Song Han, Yukang Chen

TABLE OF CONTENTS

1. News
2. Highlights
3. Introduction
4. How to contribute
5. Citation
6. License
7. Acknowledgement

News

• [2026.1.27] LongLive is accepted by ICLR-2026. 🎉🎉🎉
• [2026.1.11] Many thanks @qixinhu11 for adapting LongLive's original RoPE into KV-cache relative RoPE. Now LongLive supports generating infinite long videos!
• [2025.12.4] We fix a bug in global_sink==False mode. Now our model generate videos in higher quality.
• [2025.11.3] We implement LongLive on linear attention model SANA-Video! Now SANA-Video can generate 60s interactive videos in real-time.
• [2025.11.1] The license has been changed from CC-BY-NC-SA 4.0 to Apache 2.0.
• [2025.10.11] Many thanks to @yondonfu for building an interactive UI based on LongLive. Please check it here.
• [2025.10.1] We compare Sora2 (+ GPT-5 prompt engineering) with LongLive-1.3B in the interactive long video generation. See here for details.
• [2025.9.30] We release example prompts to reproduce our demo videos.
• [2025.9.29] We release Paper, this GitHub repo LongLive with all training and inference code, the model weight LongLive-1.3B, and demo page Website.

Highlights

1. Long Video Gen: LongLive supports up to 240s video generation, with visual consistency.
2. Real-time Inference: LongLive supports 20.7 FPS generation speed on a single H100 GPU, and 24.8 FPS with FP8 quantization with marginal quality loss.
3. Efficient Fine-tuning: LongLive extends a short-clip model to minute-long generation in 32 H100 GPU-days.

Introduction

LongLive accepts sequential user prompts and generates corresponding videos in real time, enabling user-guided long video generation.

Please see our docs for Installation, Training, and Inference.

How to contribute

• Make sure to have git installed.
• Create your own fork of the project.
• Clone the repository on your local machine, using git clone and pasting the url of this project.
• Read both the Requirements and Installation and Quick Guide sections below.
• Commit and push your changes.
• Make a pull request when finished modifying the project.

Citation

Please consider to cite our paper and this framework, if they are helpful in your research.

@article{yang2025longlive,
      title={LongLive: Real-time Interactive Long Video Generation},
      author={Shuai Yang and Wei Huang and Ruihang Chu and Yicheng Xiao and Yuyang Zhao and Xianbang Wang and Muyang Li and Enze Xie and Yingcong Chen and Yao Lu and Song Hanand Yukang Chen},
      year={2025},
      eprint={2509.22622},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}

Acknowledgement

• Self-Forcing: the codebase and algorithm we built upon. Thanks for their wonderful work.
• Wan: the base model we built upon. Thanks for their wonderful work.