What Is Wan 3.0
Wan 3.0 is Alibaba's 30-second document-to-video AI model that helps teams turn a product deck or PDF into a finished clip in a single generation. The Tongyi Wan team opened public beta on August 6, 2026 and pushed it to general availability on August 24, 2026 through Alibaba Cloud Model Studio. It is aimed at marketers, educators, and sellers who need video without a shoot.
Core Features
- Native 30-second generation: one pass produces a full narrative scene, up from 15 seconds in the prior Wan2.7 release.
- Document input: feed a doc, xls, ppt, pdf, or md file, up to 100MB and 50 pages, and the model builds the video from its content.
- Omni-Reference: combine text, image, audio, and video references, up to 20 sources, for consistent characters and props.
- Built-in editing: change visuals, plot, or dialogue on an existing clip without regenerating from scratch.
- Real-world fidelity: diverse, lifelike faces with stable micro-expressions instead of the usual same-face look.
Use Cases
- Brands converting a product PPT into a launch ad or brand film.
- Trainers turning course decks into narrated video lessons.
- Agencies producing short-form ads at scale from spreadsheet data.
Pricing
Wan 3.0 is billed per second on Alibaba Cloud Model Studio: $0.05 at 480P, $0.10 at 720P, and $0.20 at 1080P, so a 30-second 1080P clip costs about $6. A free public beta ran on the Wanxiang site and the Qianwen app. Explore related generators in our AI video category.
Our Take
Best for teams with existing office documents who want video fast; the trade-off is that prompt control is looser than Runway or Seedance, so fine directorial tweaks take more rounds. Pick it when the source material is already a deck or report.
