WAN 3.0

WAN 3.0 is for creators working from real source materials, bringing references, duration and pacing, consistency, and editing into one video workflow.

20 reference assetsDocument and web page referencesNative 30-second video generation

Use in NadouPro

Input support for 20 reference assets

WAN 3.0 accepts multiple input types, letting visuals, sound, documents, and web page context shape the video output without first converting everything into a single format.

Up to 20 referencesImagesVideosAudioDocumentsWeb pages

Native 30s generation gives full ideas more room

With native 30-second generation, WAN 3.0 gives scenes, pacing, and information structure more space to develop. Smart duration support helps the runtime match the creative direction.

30 secondsSmart durationNarrative pacing

Keep details consistent

Pixel-level consistency helps preserve important reference details, while video edits can follow your instructions or use reference assets to guide changes.

Detail preservationConsistencyInstruction-guided editingReference-guided editing

Use cases

Where WAN 3.0 fits

Product and brand videos

Use images, clips, audio, documents, or web pages as references to shape a 30-second video around existing source materials.

Story-led social content

Use the longer native duration to plan a beginning, development, and ending with clearer pacing.

Video revisions and variants

Edit with instructions or reference assets to refine video and move through multiple iterations more efficiently.

Specs

Model WAN 3.0

Platform NadouPro

Reference capacity Up to 20 reference assets can be used in one generation.

Reference input types Images, videos, audio, documents, and web pages can all be used as references.

Duration Supports native 30-second generation and includes smart duration support.

Consistency The model supports pixel-level preservation of reference details.

Video editing Edits can be guided by instructions or by reference assets.

Frequently asked questions

What references can WAN 3.0 use?

It can use images, videos, audio, documents, and web pages, with up to 20 reference assets supported in one generation.

How long can WAN 3.0 generate natively?

WAN 3.0 supports native 30-second video generation and also includes smart duration support.

How does it handle consistency?

The model supports pixel-level consistency, helping the generated video stay closer to details in the reference assets.

Can WAN 3.0 edit existing video?

Yes. Video editing can follow your instructions or be guided by reference assets.

Contains AI-generated content