Alibaba's open-source AI video generation models using Mixture-of-Experts architecture for text-to-video and image-to-video synthesis.

Features

  • Mixture-of-Experts (MoE) architecture with specialized generation stages
  • Both text-to-video and image-to-video variants
  • Bilingual prompt support (Chinese and English)
  • Spatiotemporal variational autoencoder for motion quality
  • 720p to 1080p output resolution

Wan Video is a family of open-source video generation models developed by Alibaba’s Wan-AI initiative. The Wan 2.2 series introduces Mixture-of-Experts architecture to video generation, using a high-noise expert for initial layout and a low-noise expert for detail refinement. Available in both text-to-video (T2V) and image-to-video (I2V) variants, Wan is a strong option for developers and researchers who want full control over the generation pipeline.

Similar tools