Wan AI

[Hangzhou, China][founded 2023][backed by Alibaba Group (Alibaba Cloud Tongyi Lab)]wan.video

Wan (Tongyi Wanxiang) is the visual generation line of Alibaba Cloud's Tongyi Lab, launched as an image model on 7 July 2023 and known since 2025 for open-weights video models: Wan 2.1 (weights released 25 February 2025) and Wan 2.2 (28 July 2025), both Apache 2.0 and both covering text-to-video, image-to-video and text-to-image. The catalog entries here are the text-to-image and image-to-image endpoints of Wan 2.2 and of the proprietary Wan 2.5 preview shown at Apsara Conference in September 2025; Wan 2.6 followed on 16 December 2025. Qwen language models are listed on the Qwen page.