Thirty seconds, one take
Longer runtimes generated in a single pass rather than assembled from fragments. No seams where one clip ends and the next begins, and no drift in the subject halfway through.
Why creators choose WAN 3.0
Start CreatingEvery kind of reference
Bring in up to ten images, five video clips, and five audio clips at once. Images hold the subject, video carries the movement, audio shapes the sound.
First and last frame
Define where the clip opens and where it closes, then let the model build the movement in between. The result lands where you planned it.
Sound with the picture
Audio is produced as part of the same generation rather than added later, so the clip arrives complete instead of silent.
How to create with WAN 3.0
Start creating with Alibaba’s WAN 3.0 by following a few simple steps directly inside the DaVinci AI Toolkit.
Who WAN 3.0 is for
WAN 3.0 by Alibaba suits creators who need longer, sound-complete video with real control over subject and motion.
Content creators
Fill a whole short-form slot with one generation instead of cutting three together. Keep the same character and the same setting from the first second to the last.
Agencies and brands
Produce full-length concept spots rather than fragments. Feed in product shots and a reference clip and get something close to a finished cut back.
Filmmakers and creative teams
Explore a scene at its real length. Set the opening and closing frames, hand over reference footage for the camera move, and watch the whole beat play out.
Content creators
Fill a whole short-form slot with one generation instead of cutting three together. Keep the same character and the same setting from the first second to the last.
Agencies and brands
Produce full-length concept spots rather than fragments. Feed in product shots and a reference clip and get something close to a finished cut back.
Filmmakers and creative teams
Explore a scene at its real length. Set the opening and closing frames, hand over reference footage for the camera move, and watch the whole beat play out.
Try WAN 3.0
Enter a text prompt, upload an image, or add references, and generate a video with sound in seconds.
Powerful features
WAN 3.0 brings together long single-pass generation, multimodal references, and native audio to produce video from text or images.
Text, image, and reference
Generate from a written prompt, animate a single still, or guide the result with images, video clips, and audio referenced directly in the prompt.
Extended single-pass duration
Longer clips generated in one go, so movement and identity stay consistent across the full runtime instead of resetting at every cut.
Native audio generation
Sound is created alongside the visuals as part of the same pass, covering effects, ambience, and speech.
Keyframe control
Set a starting image and an ending image and the model generates the transition between them, giving the clip a defined beginning and end.
Text, image, and reference
Generate from a written prompt, animate a single still, or guide the result with images, video clips, and audio referenced directly in the prompt.
Extended single-pass duration
Longer clips generated in one go, so movement and identity stay consistent across the full runtime instead of resetting at every cut.
Native audio generation
Sound is created alongside the visuals as part of the same pass, covering effects, ambience, and speech.
Keyframe control
Set a starting image and an ending image and the model generates the transition between them, giving the clip a defined beginning and end.
