# LongCat Video

> Meituan's video model for text, image and continuation, with an avatar variant

- Category: Image, video & voice (`media-ai`)
- Tags: `open-source`, `free`
- Official: https://github.com/meituan-longcat/LongCat-Video
- See also: https://huggingface.co/meituan-longcat/LongCat-Video-Avatar-1.5
- Wunderlust: https://wunderlust.xyz/r/longcat-video
- Image: https://wunderlust.xyz/media/longcat-video-poster.jpg
- Video preview: https://wunderlust.xyz/media/longcat-video.webm
- Added: 2026-11-05

## What it is

A 13.6-billion-parameter open source model that unifies text-to-video, image-to-video and video continuation, with LongCat-Video-Avatar models for audio-driven video.

## What it’s for

Use it to generate long videos or animate an avatar from an image and an audio track.

## Benefits

- A single architecture for three tasks, with long videos
- Avatar variant (v1.0 and 1.5) with multi-speaker audio and INT8 quantization
- Weights and code published

## Caveats

- MIT license, no pricing, but requires a CUDA GPU and a multi-step installation
- According to the authors, it delivers 720p at 30 fps with quality comparable to leading models; the tests are internal and weren't independently verified

## Similar on Wunderlust

- [Higgsfield CLI](https://wunderlust.xyz/r/higgsfield-cli.md): A CLI to generate images, video, audio and 3D with Higgsfield models from the terminal
- [Pocket TTS](https://wunderlust.xyz/r/pocket-tts.md): Kyutai's voice model that synthesizes and clones voice on CPU
- [OpenCut](https://wunderlust.xyz/r/opencut.md): Privacy-focused open source video editor, available in the browser and on GitHub
- [ComfyUI](https://wunderlust.xyz/r/comfyui.md): An open source node interface for building image, video and audio generation flows
- [VoiceStudio](https://wunderlust.xyz/r/voicestudio.md): Voice, cloning, transcription and dubbing
- [OpenMontage](https://wunderlust.xyz/r/openmontage.md): An open source video production system orchestrated by a coding agent