Wan is an open family of video generation models that turn text or images into short video clips, with code and weights free to download.
In plain words
Wan is built on a diffusion transformer, the same broad design used by many modern image and video generators. The first open release, Wan2.1, came out in February 2025 in two sizes, 1.3 billion and 14 billion parameters. The small one needs only about 8 GB of graphics memory, so it can run on many home gaming GPUs. Wan2.2 later added a mixture-of-experts design.
A quick example
Imagine typing a prompt such as "a paper boat drifting down a rainy street". A Wan text-to-video model starts from random noise and gradually shapes it into a few seconds of video that match the description.
Where you'll meet it
- Text-to-video and image-to-video generation.
- Video editing, including in tools like ComfyUI.
Related
Sources
- PaperWan: Open and Advanced Large-Scale Video Generative Models, Team Wan (arXiv) · read 29 Sept 2026
- DocsWan2.1, Wan-Video (GitHub) · read 29 Sept 2026
- DocsWan2.2, Wan-Video (GitHub) · read 29 Sept 2026
Checked by WhatIsAI against the sources above · 29 Sept 2026. Spotted an error? Report it.