Alibaba Launches Public Beta for Wan3.0 Video Generation Model
Alibaba has launched the public beta of Wan3.0, its next-generation video generation model that supports single-clip generations of up to 30 seconds. Notably, it is the first to support document formats like PDF, PPT, and Excel as inputs to directly generate videos such as product demos and tutorials. By extending video length to 30 seconds and allowing structured documents as inputs, Wan3.0 bridges the gap between static office files and dynamic video content. This significantly enhances productivity for creators, educators, and businesses looking to automate video production from existing documents. The model accepts document files up to 100MB and 50 pages, and features high consistency in characters, props, and styles to avoid visual discrepancies. API pricing is set at 0.3, 0.6, and 1.2 RMB per second for 480P, 720P, and 1080P resolutions, respectively.
## BACKGROUND
Traditional AI video generators are typically limited to short clips of 4 to 15 seconds, making it difficult to tell a coherent story in a single generation. Additionally, most models only accept text prompts or images as inputs, requiring manual conversion of documents into prompts before video generation can occur.