Wan is a family of open-source video generation models and an AI creative platform developed by Alibaba's Tongyi Lab. The platform at wan.video supports text-to-video, image-to-video, text-to-image, image-to-image, and image editing.
Wan2.2, released under the Apache 2.0 license, generates video at 480p and 720p from text or images. It uses a Mixture-of-Experts architecture with two denoising experts, totaling up to 27 billion parameters with 14 billion active. The compact TI2V-5B unified model can generate a 5-second 720p video in under 9 minutes on a single consumer GPU. The family also includes Speech-to-Video (S2V) and Wan-Animate for character animation and replacement. Model weights are published on Hugging Face and ModelScope, and the models have a large following in the open-source and ComfyUI communities.
Newer proprietary versions (Wan 2.5 and later, with native audio generation) are served through the web platform and Alibaba Cloud. API access is available via Alibaba Cloud Model Studio.
The open-source models are free to self-host under Apache 2.0. The wan.video platform offers free credits plus paid subscription tiers and one-time credit packs with commercial licensing included.