Panify
Wan model family

Wan AI Models

Explore Alibaba's Wan family—from open video generation foundations to longer multimodal creation, native audio and precision editing.

Latest capability set

Longer, reference-rich creation
Wan 3.0

Native 30-second generation

Images, video and audio references

Immersive stereo audio

Precision editing and extension

  • Text, image and reference-led video
  • Native audio-video in newer models
  • Generation, editing and extension
  • Developed by Alibaba's Wan team

Model overview

What is Wan?

A generative model family spanning video, image, audio and editing workflows.

Wan is Alibaba's family of generative video and image models. The series covers text-to-video and image-to-video creation and has expanded into keyframe control, reference-to-video, character animation, native audio-video generation and instruction-led editing.

The newest official Wan experience brings generation and editing into a broader multimodal workflow. Capabilities, model IDs, deployment regions and output limits still vary by endpoint, so Panify verifies each workflow before making it available.

Creative system

From a brief to a controlled audio-visual result

Choose the Wan workflow around the source material and control your project needs.

Text and image generation
Create video from a written brief or animate an authorized starting image with directed action and camera movement.
Reference-led consistency
Use supported images, videos and audio to guide characters, props, scenes, voices, motion and spatial relationships.
Native audio-video
Newer Wan workflows coordinate dialogue, ambience, music and visual rhythm in one generated result.
Generation and editing
Create a new scene or refine existing video through instructions, references, targeted changes and video extension.

Model timeline

The evolution of Wan

Wan has grown from open visual generation into managed, multimodal production workflows.

Wan 2.1
Open foundation

Open text-to-video and image-to-video models, including first-and-last-frame control.

Wan 2.2
Expanded workflows

Added open variants and specialized character animation and replacement workflows.

Wan 2.5
Native audio

Introduced managed synchronized audio-video generation across text and image workflows.

Wan 2.6
Multi-shot

Expanded audio-enabled multi-shot and reference-to-video creation.

Wan 2.7
Documented API

Supports 2–15 second output up to 1080P, custom audio and reference or editing endpoints.

Wan 3.0
Latest

Highlights native 30-second creation, richer multimodal references, precise editing, consistency and stereo audio.

Explore this model

Creation modes

Start with the material that communicates your intent

Each workflow gives a different balance of speed, reference control and editability.

Text to video
Direct cinematic shots, complex action and narrative progression with a chronological brief.
Image to video
Animate a starting image into a longer audio-video sequence with automatic shot planning.
Reference to video
Use an authorized person or object as the recurring subject for solo or multi-character scenes.
Video editing
Change visuals, plot, environment, lighting or sound, or extend a supported video.

Creative applications

What can you create with Wan?

Use the family for both early creative exploration and more controlled production concepts.

01

Advertising concepts

Develop longer product and campaign narratives with controlled shot progression.

02

Character scenes

Plan solo or ensemble performance with authorized identity and voice references.

03

Cinematic action

Previsualize movement, camera language and physically complex scenes.

04

Product storytelling

Preserve important product details across connected shots and settings.

05

Music and dialogue

Build scenes around speech, ambience, rhythm and stereo sound.

06

Video revision

Refine subjects, backgrounds, lighting or sound and extend supported footage.

Wan FAQ

Frequently asked questions

Who developed Wan?

Wan was developed by Alibaba's Wan team and is available through official Wan experiences, open-model channels and Alibaba Cloud Model Studio, depending on the version.

What can Wan generate?

Depending on the version, Wan supports text-to-video, image-to-video, reference-to-video, native audio, character animation, image generation and video editing.

Does Wan generate audio?

Supported newer workflows include native audio-video generation. The current official product highlights stereo sound, natural timbre, dialects, music and coordinated audio-visual rhythm.

Is Wan open source?

Some Wan generations and specialized models have open weights, while newer managed capabilities may use hosted endpoints. Check the exact model's official license.

Which Wan version should I choose?

Choose around the required inputs, audio, duration, editing controls, deployment region and cost. Panify lists only verified workflows and settings.

Explore the next Wan creative workflow

Review Wan 3.0's official capability set and prepare your authorized references while Panify verifies production access.