dGENdGEN Visual Studio
← All AI modelsby Alibaba

Wan 3.0

Start and end frames, its own audio, and motion that stays coherent.

Open it in the StudioExplore all AI models

Clip made with Wan 3.0 in this studio.

What it is

Alibaba describes Wan 3.0 as "the latest generation AI video model, delivering enhanced motion smoothness, superior scene fidelity, and greater visual coherence". Those three claims are about one problem: video models drift over time, and this generation is tuned against that.

Unlike the 2.7 line it takes both a start frame and an end frame, so you can decide where a shot begins and where it lands instead of only where it starts.

It generates its own audio rather than asking you to supply a track, and it runs at 480p, 720p and 1080p.

Prime is the same model at a different speed and price. Which one you get is a choice in the studio, not a different page.

Made with Wan 3.0 in this studio

Clip

What it is good at

Both ends of the shot

Approve a first and a last frame and it moves between them. That is the difference between describing a shot and directing one.

Sound made with the picture

Audio is generated alongside the video rather than laid over it afterwards, so what you hear belongs to what you see.

One subject across shots

The reference ingress takes images, clips and audio together, which is what you need when the same character has to come back in the next shot.

Where it struggles

  • Alibaba publishes no list of limitations for this model, so none is invented here.
  • The reference ingress and the frame ingress are separate endpoints. You pick one per shot; the studio switches for you when you attach something the current one cannot read.

How to prompt it

Describe one continuous action and let the scene stay still around it. Coherence is what this generation is tuned for, so one thing to track beats three simultaneous events.

An example that shows what that means

A single red kite climbing steadily against a pale sky, the camera tilting up to follow it, thin cloud moving slowly behind, the line taut and visible against the light, nothing else in frame.

What we offer here

Straight from the catalogue, so this list never goes stale. The price is the provider’s published rate with our markup on top, and it is on the button before you run anything.

Wan 3.0 · from textVideofrom 2,000 cr
Wan 3.0Videofrom 2,000 cr
Wan 3.0 PrimeVideofrom 2,800 cr
Wan 3.0 Prime · from textVideofrom 2,800 cr
Wan 3.0 · from a referenceVideofrom 2,000 cr
Wan 3.0 Prime · from a referenceVideofrom 2,800 cr

Questions

What is the difference with Wan 2.7?

Wan 3.0 takes an end frame as well as a start frame, and it generates its own audio. On 2.7 you supply the audio and only the start is yours to choose.

And what does Prime change?

Speed and price, not what it can do. Both stands have the same inputs and the same resolutions; the studio shows what each one costs before you run it.

Can I keep the same character across several shots?

Yes, through the reference ingress, which takes images, clips and audio together. Attach a reference and the studio moves you to that variant.

Alternatives

SeedanceMulti-shot sequences with big movementKling 4.0Long clips of up to 30 seconds, told in several shotsKling 3.0Cinematic clips up to 15 seconds with a recurring characterMiniMax H3 MaxRecasting, extending and lip-syncing footage you already have

Guides with Wan 3.0

Start and end framesHand a model both ends of a movement and it fills in the middle. Here is when that beats a prompt.A music video for your own trackYour artist performs your track, close up and in motion, with the mouth on every word.The best image to video AI for your photoTurning a photo into a moving shot: which AI model holds your image best, steers the ending, or tests ideas cheaply.

Next

Smooth transitionsTwo stills in, the move between them out.Which model for which jobA plain comparison, no scores.

Sources

Try Wan 3.0 in the studio

Compare them on your own prompt instead of on a demo. You start with 1,000 free credits.

Get early access