Veo 3.1
Generate short videos with native audio from text prompts, images and visual references.
Wan 3.0
www.alibabacloud.com
Alibaba’s video generation model turns prompts, media references, documents and public webpages into short videos with native audio.
Visit websitePricing
Paid
Available on
Alibaba Cloud Model Studio API
Category
AI Video Tools
What is Wan 3.0?
Wan 3.0 is Alibaba’s video generation model for creators, production teams and developers. It accepts text, reference media, documents and public webpages, supporting filmmaking, short dramas, social content, marketing and educational videos. These inputs let teams develop moving scenes from existing briefs and creative assets.
Alibaba Cloud Model Studio provides Standard and Prime variants through an API that is currently in preview. Generation runs as asynchronous tasks: applications submit a request and retrieve the result after processing. This profile covers the Model Studio offering and its usage-based billing.
Key features
Multiple creation modes
Generate from text, animate a first-frame image, specify first and last frames, or use multimodal references to guide a new video.
Media and document references
Use up to 20 reference materials total, subject to limits of 10 images, five videos and five audio clips. One document or public webpage also counts toward that total.
Native audiovisual generation
Create videos up to 30 seconds long at 30 FPS, with generated dialogue, background music and sound effects.
Video editing and extension
Use reference footage to add, remove or change elements, revise lighting or dialogue, transform style, or extend a sequence.
Output controls
Select 480P, 720P or 1080P output, choose an aspect ratio, request automatic duration selection, and enable or disable audio.
Who is it for?
Marketing teams exploring campaign scenes
Develop short video concepts using product images and existing campaign material as references.
Educators preparing visual explainers
Explore short explanations based on presentation material or a public information page.
Filmmakers revising a shot
Try changes to an existing scene’s lighting, objects or dialogue before choosing a creative direction.
How to use Wan 3.0
- Review the Model Studio guide and select a Wan 3.0 variant available in your region.
- Configure a workspace and API key with a matching regional endpoint.
- Describe the scene and supply compatible frame, reference, document or webpage inputs.
- Choose output settings, submit a generation task and poll its status.
- Review the result and download the video before its link expires.
Wan 3.0 pricing
Paid
Model Studio API usage is charged by video duration, with rates depending on region, resolution and Standard or Prime variant. Both reference-video input and successful output contribute to billing; failed requests are not charged. Singapore offers a limited introductory quota. Check official pricing for current regional rates.
Check current pricing ↗What to know before you use it
- When video references are supplied, their combined duration plus the generated output cannot exceed 30 seconds.
- Frame-based modes cannot be mixed with reference-mode assets. A document and a webpage cannot be supplied together.
- Reference videos total at most 15 seconds, and reference audio separately totals at most 15 seconds. Webpage inputs must be publicly accessible without login.
- The model, endpoint and API key must belong to the same region. Generated download links expire after 24 hours.
Frequently asked questions
Can Wan 3.0 generate silent video?
Yes. Disable audio in the output settings to generate a video without an audio track.
Can I control the beginning and ending images?
Yes. Supply first and last frames using the dedicated frame mode to guide the transition between them.
Does the introductory quota apply everywhere?
No. The 30-second combined input and output quota applies only in Singapore and is valid for 90 days from the applicable activation, release or approval date.