Alibaba's Wan3.0 video model goes generally available, adding PDF and slide-deck inputs.
The AMW Read
GA release of an existing Alibaba video model with expanded document inputs is an incremental capability update, not a new entrant or debate-resolving event.
Alibaba's Wan3.0 video model goes generally available, adding PDF and slide-deck inputs.
Alibaba has moved its Wan3.0 video generation model to general availability, positioning it to produce 30-second clips from text, image, video, or audio prompts. Beyond standard prompt formats, the model accepts structured document inputs — web pages, PDFs, and slide decks — as source material for video generation. Promotional material for the launch included an AI-generated clip featuring actor Tom Cruise's likeness, drawing renewed scrutiny over unauthorized use of a celebrity's image in generative output.
The release extends Alibaba's presence in the generative video race, where Chinese and US labs are both pushing multi-modal input handling as a differentiator beyond raw clip fidelity. Accepting documents and slide decks as generation inputs signals an effort to pull video generation toward business content workflows — turning existing static materials into video rather than requiring a written prompt each time. The Tom Cruise likeness controversy is also the latest instance of a recurring problem across generative media: promotional demos built on recognizable public figures routinely outrun the rights clearance needed to use their image, exposing labs to takedown and legal risk even as the underlying technology ships.
For builders integrating generative video into marketing or content pipelines, Wan3.0's document-to-video input is a concrete new entry point worth testing against existing prompt-only tools, but any workflow touching real people's likeness should assume no implicit rights clearance from the model provider. Investors tracking the generative video field should treat input-modality breadth, not just clip length or resolution, as the near-term competitive axis among Alibaba, and other frontier video labs.
