Alibaba / Video model

Wan 3.0

Alibaba's long-form video model with native audio, clips from 2 to 30 seconds, and up to four reference images for consistent subjects.

Meet Wan 3.0

Wan 3.0 is Alibaba's video model. It generates clips from 2 to 30 seconds from text, a start and end frame, or up to four reference images, with native audio that you can switch off.

It is one of the cheaper routes to 1080p: unlike most models here, 1080p costs only twice the 720p rate.

Available in Advibly

Settings and estimated costs from Advibly's live model configuration.

Inputs
Text, Start frame, End frame, Up to 4 reference images
Durations
2s to 30s
Aspect ratios
9:16, 16:9, 4:3, 3:4, 1:1
Output options
480p, 720p, 1080p
Audio
Generated with the video, can be switched off
Estimated cost
$0.08 to $0.30 / secondOne credit costs $1. The range spans the cheapest and the most expensive settings this model offers; the composer shows the exact amount before you generate.0.08 to 0.3 credits / second

When to use it

Use Wan 3.0 for longer 1080p shots where the subject has to stay the same.

  • Multi-reference subject consistency
  • Longer 1080p clips at a moderate price
  • Silent b-roll when you have your own audio

Things to keep in mind

21:9 is not available.

Switch audio off when you will replace it anyway.

Get more from your prompt

  1. 1.Give each reference a clear role in the prompt.
  2. 2.For clips past 15 seconds, write the action as timed beats.
  3. 3.Describe the camera path across the whole clip, not only the opening.
  4. 4.Match the lighting of your reference images in the prompt.

Common questions

What is Wan 3.0 best used for?

Wan 3.0 is particularly useful for multi-reference subject consistency, longer 1080p clips at a moderate price, silent b-roll when you have your own audio. The right choice still depends on your input assets, delivery format, turnaround and credit budget.

What inputs does Wan 3.0 support in Advibly?

Wan 3.0 works from text, start frame, end frame and up to 4 reference images. Advibly only shows the controls a model actually supports once it is selected, so you are never offered a combination it cannot render.

How many credits does Wan 3.0 use?

The current estimate is 0.08 to 0.3 credits / second ($0.08 to $0.30 / second, at $1 per credit). The exact amount follows the options you pick, and Advibly shows it on the generate button before anything is charged.

Reviewed September 29, 2026 against Advibly's model configuration and Alibaba's published material.

Developer source

Create with Wan 3.0

Open a new video workspace with this model selected.

Create with Wan 3.0
Back to video models