🎉 Wan 3.0 is live — Alibaba's newest video model. Native 30-second clips at 1080P with synced audio, guided by up to 20 image, video, and audio references!
Wan 3.0 Video Generator
Generate videos using cutting-edge AI models including Veo 3, Sora 2, and more - with optional reference images
Wan 3.0Note: Images will be used as reference materials to guide the video generation.
Upload up to 5 reference audio clips (total ≤15s) to guide audio generation.
✨ Please login to get free credits ✨
Video Reframe
Change the aspect ratio of any video up to 30 seconds long
Click to upload or drag and drop
Formats: MP4, WebM, QuickTime
✨ Please login to get free credits ✨
Wan 3.0 — 30-Second AI Video with Native Audio and Omni-Reference
How to Use Wan 3.0 on LuminaMind
Turn a prompt and a few references into a 30-second, audio-synced clip with Wan 3.0.
Why Choose Wan 3.0?
Alibaba's most capable video model yet—built for longer stories, richer references, and audio that arrives with the picture.
Native 30-Second Generation
Wan 3.0 doubles the 15-second ceiling of Wan 2.7. Generate a complete 30-second scene with multi-shot narrative and director-level camera moves in one pass—no stitching.
Omni-Reference Inputs
Combine up to 10 images, 5 video clips, and 5 audio tracks in a single generation to lock in characters, props, locations, and voices. Documents and web pages can be references too.
Pixel-Perfect Consistency
Faces, wardrobe, and products stay recognizable from the first shot to the last, with fewer deformations and clipping artifacts than earlier Wan releases.
Native Synced Audio
Dialogue, background music, and sound effects are generated together with the video, so lips, beats, and impacts land on the right frame without post-production dubbing.
Reality-Grade Rendering
Sharper hair, glass reflections, water, and lighting, plus more expressive faces and micro-expressions for performances that read as real.
Precision Editing & Text
Edit generated clips by instruction or reference, extend them forward or backward, and render legible on-screen text, charts, and formulas in 12 languages.
Wan 3.0 at a Glance
Longer, more controllable, audio-native
Max Length
30s
Single pass
Resolution
1080P
30 fps
Native Audio
Synced
Dialogue, music, SFX
Where Wan 3.0 Shines
Official Wan 3.0 showcases—single-take rides, surreal commercials, comedy, cinematic realism, music, and sci-fi action.
30-Second Single-Take Ride
A kid rides a shopping cart down a palm-lined Los Angeles boulevard, weaving through traffic and launching skyward, all in one unbroken 30-second shot with the camera locked on the action.
Fantasy Product Commercial
A white-haired girl opens the fridge and shrinks into a frozen world of juice cartons, jelly monsters, and broccoli forests, then returns with dessert. One continuous clip, five surreal set pieces.
Deadpan Desert Comedy
A Route 66 mechanic lounges outside a sun-bleached gas station, toothpick in mouth, as a cowgirl rides up on horseback. Dry timing, natural light, and expressive close-ups carry the joke.
Cinematic Realism
A woman in a crimson gown stands in a misty river as a herd of black horses charges past her. Splashing water, flowing fabric, and muscle motion all hold up under a slow cinematic push-in.
Music & Performance
A violinist plays through pouring rain under cold teal light, water streaming off the instrument. Bow strokes stay in sync with the generated score for a moody, music-video-ready single take.
Sci-Fi Action Sequence
Combat robots battle through a burning city street while a single white flower survives in the rubble. Explosions, debris, and metal impacts land with synced sound across the full 30 seconds.
Pricing
Credits can be used for video generation with multiple AI models including Wan 3.0 and more.
700 Credits
Most popular for individual creators!
Includes
- 700 credits / month
- Credits never expire
- 4K Video Resolution
- Text/Image/Video to Video:
Veo 3.1
Sora 2
Seedance 2.5Wan 3.0 - Text/Image to Image:
GPT Image 2
Nano Banana 2 - No Watermark
- Private Generation
- Commercial License
cancel anytime
400 Credits
Perfect for trying out.
Includes
- 400 credits / month
- Credits never expire
- 4K Video Resolution
- Text/Image/Video to Video:Veo 3.1Sora 2Seedance 2.5Wan 3.0
- Text/Image to Image:GPT Image 2Nano Banana 2
- No Watermark
- Private Generation
- Commercial License
cancel anytime
1500 Credits
Best for professional creators!
Includes
- 1500 credits / month
- Credits never expire
- 4K Video Resolution
- Text/Image/Video to Video:Veo 3.1Sora 2Seedance 2.5Wan 3.0
- Text/Image to Image:GPT Image 2Nano Banana 2
- No Watermark
- Private Generation
- Commercial License
- Priority Support
cancel anytime
Made with Wan 3.0
See what creators are making with Wan 3.0
The Technology Behind Wan 3.0
How Alibaba's unified multimodal transformer reaches 30 seconds.
One DiT for Every Task
Text-to-video, keyframes, references, editing, and extension run through one flow-matching diffusion transformer. A router reads media types and prompt intent, and the whole 30-second latent is denoised as one sequence.
Joint Audio-Video Latents
Frames go through the 3D causal Wan-VAE and sound through a 1D audio VAE; both token streams are denoised together in one hybrid-MMDiT sequence, so dialogue, music, and effects emerge with the picture, not as a later dub.
Reference Token Binding
Up to 20 images, clips, and audio tracks enter as numbered reference tokens (Image 1, Video 1, Audio 1) that every frame attends to, while a thinking stage rewrites the prompt or a document into a timestamped shot script.
Wan 3.0 FAQ
Common questions about Wan 3.0
What is Wan 3.0?
Wan 3.0 is Alibaba's latest all-in-one AI video model from Tongyi Lab, released in public beta on August 6, 2026 and generally available since August 24, 2026. It generates video and native audio together from text, image, video, audio, and even document inputs, in clips up to 30 seconds.
How is Wan 3.0 different from Wan 2.5?
Wan 3.0 raises the single-pass limit from 10 seconds in Wan 2.5 to 30 seconds, merges reference-to-video, keyframes, editing, and extension into one model, accepts up to 20 multimodal references, and delivers steadier characters, richer physical detail, and cleaner audio-visual sync.
What resolution and length does Wan 3.0 support?
Wan 3.0 outputs 480P, 720P, or 1080P at 30 fps in 16:9, 9:16, 1:1, 4:3, 3:4, or adaptive aspect ratios, with any length from 2 to 30 seconds in a single generation. When you add a reference video, the input and output durations together can add up to 30 seconds in total.
Does Wan 3.0 generate audio and lip-sync?
Yes. Wan 3.0 produces dialogue, background music, and sound effects in the same pass as the picture, so speech stays lip-synced and effects land on the right frame. Put spoken lines in quotes in your prompt, or supply up to five reference audio clips to drive voices and rhythm.
What inputs can I use with Wan 3.0?
Text prompts, first and last frames, reference images, reference video clips, and reference audio can all guide one generation. Wan 3.0 takes up to 20 references in total—10 images, 5 videos, and 5 audio clips—and the model can also read documents, spreadsheets, slides, and web pages.
Is Wan 3.0 suitable for commercial use?
Yes. Native 30-second clips, 1080P output, precise multi-reference control, and synced audio make Wan 3.0 a strong fit for advertising, product videos, short drama, tourism marketing, and music videos, from quick concept tests to polished, brand-ready final deliverables.
Start Creating with Wan 3.0
Generate 30-second, audio-synced AI videos with Wan 3.0 today.