Build a flexible, developer-friendly video generation API that handles diverse input types and scales to long-form content, driving Alibaba Cloud adoption.
Wan 3.0 public beta August 6; generally available August 24. Generates up to 30 seconds per video.
Updated 24 Aug 2026Checked 11 Oct0 updates this week
Wan 3.0 accepts text, images, audio, video, PDFs, spreadsheets, slides and webpages as input, enabling video creation from diverse source material.
Wan 3.0 doubles the maximum video length to 30 seconds, enabling longer scenes and narratives compared to Wan 2.7's 15-second limit.
Video is generated as one continuous take rather than stitched fragments, improving temporal coherence and visual flow.
Wan 3.0 is a closed, API-only model available through Alibaba Cloud Model Studio and Qwen Cloud, with no open weights published.
| Spec | Wan video models |
|---|---|
| Wan 3.0 maximum video duration | 30 secondsR (reported) |
| Wan 2.7 maximum video duration | 15 secondsR (reported) |
| Wan 3.0 resolution options | 480p, 720p, 1080pR (reported) |
| Omni-Reference inputs | Text, image, audio, video, documents, slides, PDFsR (reported) |
| Generation approach | Continuous single take (not stitched fragments)R (reported) |
R reported by the company
Alibaba Qwen
Qwen open-weight models, Qwen app
Qwen is Alibaba Cloud's AI model family and one of the most used open-weight lines, with hundreds of thousands of spin-off models online. Sizes range from under 1 billion to over 1 trillion parameters.
