Alibaba launches Wan3.0 video model, generates 30s video and accepts document inputs

On August 24, Alibaba officially launched Wan3.0, its next-generation video generation model. The new version generates up to 30-second videos and for the first time supports doc, xls, ppt, pdf, and md document format inputs, allowing continuous video generation from text, images, and document materials. The model focuses on improving consistency of characters, vehicles, props, sound, and spatial relationships in long shots.

Alibaba states that during public beta, Wan3.0 has entered production workflows for short dramas, advertising, cultural tourism, and music videos, and has been integrated with creation platforms including Flova and LibTV. Wan3.0 is now available on Alibaba Cloud Bailian, Qwen AI platform, Wanxiang official website, and the Qwen App, with API access also opened. Wan3.0 supports up to 1080p video output.

Wan3.0 pricing starts at approximately $0.05/second for 480p, $0.10/second for 720p, and $0.20/second for 1080p. This is the first AI product released after Alibaba completed its HK$80 billion share placement.