Alibaba's Wan 3.0 goes generally available, turning documents into 30-second video
The model produces 30-second 1080p clips with audio in a single pass. Its distinguishing feature is Omni-Reference: documents, spreadsheets and slide decks count as inputs alongside text and images.
Alibaba has announced that its video generation model Wan 3.0 is generally available. The model can be called through Alibaba Cloud Model Studio and Qwen Cloud.
The real change in this release is duration and the range of inputs. Wan 3.0 produces 30-second clips at 30 frames per second and up to 1080p in a single pass, with audio generated in the same pass. In earlier generations video and audio were separate steps.
The distinguishing feature is Omni-Reference. Alongside text, image, audio and reference video, documents can be given as input: PDF, Word, PowerPoint, Excel, plain text, Markdown and Apple's Keynote, Pages and Numbers formats are supported, with paged formats accepted up to 50 pages. A deck or a report can therefore be turned into video directly.
The company reports improvements in instruction following, consistency across shots and audio quality.
One inconsistency in availability is worth noting: the Model Studio page presents the API as generally available and callable without an application process, while the English API reference updated on 24 August still labels the model as preview. The service is documented for the Singapore, Beijing and US East (Virginia) regions. Pricing starts at $0.05 per second for 480p output, with separate 720p and 1080p tiers.
Details are in Alibaba Cloud's post.
