Alibaba Cloud officially launched Wan3.0 on August 24, 2026. The update is not only about turning text into images: it can understand documents, spreadsheets, slides and web pages, then organize their information into a video. Alibaba says Wan3.0 can generate clips up to 30 seconds and has been used in short dramas, films, advertising, tourism promotion and music videos.
What changed in Wan3.0?
Many AI video tools ask for prompts, reference images or storyboard descriptions. Wan3.0 expands the input toward working documents: users can provide a presentation, a project brief or a web page and ask the model to build a video from it, reducing the back-and-forth between reading, scripting and writing visual prompts.
The 30-second ceiling makes it more suitable for a complete short clip rather than only a few seconds of motion. It does not mean every output is ready to publish: continuity, character consistency, text accuracy and information selection still need human review. But for proposals, social clips and product concepts, the input is closer to a real production workflow.
Understanding documents, not just a single prompt
Document input lets a video start from information that already exists. A travel team could turn an itinerary and attraction notes into a promotional draft; a brand team could turn a product presentation into a visual narrative; an education or enterprise team could turn a briefing document into watchable content. Wan3.0 is therefore about converting information into a script, rhythm and shot choices, not only generating images.
Being able to read a document does not mean the model will correctly understand every table or constraint. Users still need to check numbers, product specifications, dates, proper names and disclaimers. A natural-looking video is not a substitute for fact-checking when the content is commercial or public information.
How can people use it now?
Wan3.0 entered public beta on August 6, followed by Alibaba Cloud’s official launch announcement on August 24. The available models, application process, regional eligibility, pricing and API limits should be checked in the latest Alibaba Cloud Model Studio page; different access points may not expose exactly the same features at the same time.
For creators, the most practical use is as a pre-production tool: use documents to establish a script and visual direction, then review shots, correct text and regenerate where needed. It can shorten the path from source material to a first cut, but it does not replace planning, editing, brand review or rights clearance.
What it means for AI video
Wan3.0 suggests that generative video is moving from “who can make the prettiest single shot?” toward “who can turn existing information into usable content?” Once documents and slides become inputs, adoption is no longer limited to video creators; marketing, education, travel and product teams become potential users. The important questions are whether outputs are reliable, whether rights are clear and whether teams can place generation inside an existing review process.
For official details, see Alibaba Cloud Model Studio’s Wan3.0 page 與 the official Wan3.0 GitHub。
