Wan 3.0 AI video generator
Standard and prime video generation with multi-modal references
- Image, video and audio references
- Native audio
- First and last frame
JPG, PNG, WEBP, MP4, MOV, MP3 or WAV. Videos and audio 2 to 15 seconds.
Mention files in the prompt as @Image1, @Video1 or @Audio1.
- Give each file one clear job: a face, an outfit, a location or a camera move.
- Mention files in order, like @Image1 and @Video1, so the model knows what to copy from where.
- Reference videos and audio work best between 2 and 15 seconds.
- Audio needs at least one image or video in the stack.
The camera slowly zooms in on the figure, who turns and walks deeper into the shadows, heightening the suspense.
The camera gently glides over the lake, reflecting the beauty and stillness of the natural landscape, evoking peace.
The camera slowly pans upwards as the city awakens, capturing the gradual shift from artificial to natural light, evoking tranquility and wonder.
The camera tilts down to the steaming coffee as rays of sunlight dance around, creating an inviting and warm atmosphere.
The camera circles the DJ as the lights pulse to the beat, capturing the electrifying, energetic ambiance of the dance floor.
The gadget rotates slowly, highlighting its cutting-edge design and technology while the camera gently zooms in to focus on its features.
The camera shakes slightly as hands organize the desk, placing items neatly and opening the notebook, creating an inspiring and relatable scene.
Text overlays appear with tips as the camera smoothly pans across the plants, enhancing the educational and calm tone.
The colors pulse and shift rhythmically as the text pops in and out, creating an exciting and inviting introduction.
The camera glides through the forest, revealing hidden creatures and ancient ruins, building excitement and mystery.
The camera does quick cuts between different poses, capturing the energy and flair of urban fashion.
The camera dramatically swoops over the armies, building anticipation with the swell of the storm and distant battle cries.
What is Wan 3.0?
Wan 3.0 is the bold, wide-ranging video model on NovaMotions AI. It works from text, from a first and last frame, or from up to 9 images, 3 clips and 3 audio files, and renders 5, 10 or 15 second clips in 720p or 1080p. It shines on stylized looks, surreal ideas and anything that needs a strong visual voice.
- longest clip
- 15 s
- highest resolution
- 1080p
- reference files per clip, images, clips and audio mixed
- 12
What Wan 3.0 does well
The strengths worth knowing before you write your first prompt.
Wide creative range
From painterly fantasy to gritty documentary, Wan 3.0 commits to a style. Name the look you want, like claymation or 16mm film, and it leans in.
Multi-modal references
Combine up to 9 images, 3 video clips and 3 audio files. Borrow a palette from one image, a character from another and the motion from a clip.
First and last frame
Set where the shot begins and ends with two images. Wan 3.0 invents the journey between them, great for morphs and scene transitions.
Native audio on demand
Switch on audio for dialogue, effects and ambience rendered with the picture. Leave it off for silent loops or when you score the edit yourself.
5, 10 or 15 seconds
Pick a short beat, a full moment or a longer scene, in 720p or 1080p, in any of six aspect ratios including 21:9.
What you can make with Wan 3.0
Stylized music visuals
Turn a track into a looping visualizer or a surreal performance clip, with the song as @Audio1 setting the mood and the pace.
Concept and mood films
Pitch a world before it is built. Wan 3.0 makes art-directed mood pieces that sell a tone faster than a deck of stills.
Scroll-stopping social art
Unusual transformations, impossible camera moves and strong styles stand out in a feed. Render 9:16 and let the look do the talking.
How to use Wan 3.0
Four steps from an idea to a finished clip, all in the workspace above.
- 01
Choose how to start
Pick Media to build from reference files, Image to animate a first frame, or Text to describe the whole clip in words. You can switch modes without losing your prompt.
- 02
Add frames or references
In Media mode, upload up to 9 images, 3 video clips and 3 audio files and give each file one clear job. In Image mode, add a first frame and, if you like, a last frame. Text mode needs nothing.
- 03
Write the prompt
Describe the subject, the action, the setting, the camera move and the mood. Point to your files with @Image1 or @Video1 and put spoken lines in double quotes.
- 04
Set the shot and generate
Choose 5, 10 or 15 seconds, 720p or 1080p, an aspect ratio and audio, check the credit cost on the button, then generate. Keep the best take or extend it with Continue from last frame.
Wan 3.0 vs Wan 2.7 vs Seedance 2.0 Pro
All three accept text, frames and media references. Wan 3.0 is the most expressive, Wan 2.7 is the dependable all-rounder with a 480p draft tier, and Seedance 2.0 Pro is the pick for exact faces and products.
| Compare | This modelWan 3.0 | Wan 2.7 | Seedance 2.0 Pro |
|---|---|---|---|
| Best for | Bold styles and surreal ideas | All-round work with cheap drafts | Exact characters and products |
| Ways to create | Media, Image, Text | Media, Image, Text | Media, Image, Text |
| Longest clip | 15 s | 15 s | 15 s |
| Highest resolution | 1080p | 1080p | 4K |
| Aspect ratios | 16:9, 9:16, 1:1, 4:3, 3:4, 21:9 | 16:9, 9:16, 1:1, 4:3, 3:4, 21:9 | 16:9, 9:16, 1:1, 4:3, 3:4, 21:9 |
| Media references | Up to 9 images, 3 video clips, and 3 audio files | Up to 9 images, 3 video clips, and 3 audio files | Up to 9 images, 3 video clips, and 3 audio files |
| Native audio | Optional | Optional | On by default |
| Typical render time | About 3 minutes | About 3 minutes | About 3 minutes |
| Starting price | From 46 credits | From 35 credits | From 35 credits |
Wan 3.0 specs
These limits come straight from the workspace settings, so they always match what you can pick above.
- Ways to create
- Media references, Image to video, and Text to video
- Clip lengths
- 5, 10, or 15 seconds
- Resolutions
- 720p or 1080p
- Aspect ratios
- 16:9, 9:16, 1:1, 4:3, 3:4, or 21:9
- Image input
- First frame plus an optional last frame
- Media references
- Up to 9 images, 3 video clips, and 3 audio files
- Native audio
- Yes, switch it on per clip
- Prompt length
- Up to 10,000 characters
- Takes per run
- 1 to 4
- Typical render time
- About 3 minutes
- Starting price
- From 46 credits
Prompt examples for Wan 3.0
Press Try this and the prompt lands in the workspace above, ready to generate or tweak.
- Text to video
Paper city wakes up
A city made entirely of folded paper unfolds at sunrise, buildings popping up like a pop-up book while paper birds take flight. Slow crane up from street level, soft warm light through translucent paper, gentle rustling and wind chimes.
- Text to video
Desert noir chase
Grainy 16mm look. A vintage car speeds down an empty desert highway at dusk, dust trailing behind, a motorcycle closing in. Low angle tracking shot from the roadside, orange and teal grade, engine roar and a twangy guitar sting.
- Media references
Painted portrait comes alive
The woman from @Image1 steps out of an oil painting in the style of @Image2, brush strokes trailing from her dress as she walks through a candlelit gallery. Slow orbit around her, warm flicker, soft footsteps and a distant cello.
Try thisUpload your files in Media mode first
Wan 3.0 questions
Short answers about modes, limits, credits and rights.
Ready to try Wan 3.0?
The workspace is at the top of this page. Generations start from 46 credits, and failed jobs return their credits automatically.
Still have a question?
Our team answers within one business day. The prompt guide covers references, camera moves, text and sound.
What is Wan 3.0?
Wan 3.0 is a video model on NovaMotions AI known for its wide creative range. It makes 5, 10 or 15 second clips in 720p or 1080p from text, from a first and last frame, or from reference images, clips and audio.
What is Wan 3.0 best at?
Stylized and imaginative work. It handles strong art directions like stop motion, watercolor, film grain or surreal transformations, and it is happy to try ideas that more literal models play safe with.
Can Wan 3.0 use references?
Yes. In Media mode you can add up to 9 images, 3 video clips and 3 audio files and mention them as @Image1, @Video1 or @Audio1. Use images for characters and style, clips for motion and audio for rhythm.
Which clip lengths and resolutions can I choose?
Clips are 5, 10 or 15 seconds long, in 720p or 1080p. All six aspect ratios are available: 16:9, 9:16, 1:1, 4:3, 3:4 and 21:9.
Does Wan 3.0 have sound?
It can generate native audio with dialogue, effects and ambience. Audio starts switched off, so turn it on for clips that need sound. It adds a few credits to the clip.
How do I get a consistent style across clips?
Reuse the same style reference image as @Image1 in every prompt and repeat the same style words, such as hand-painted gouache with visible paper texture. Keeping the seed helps when you only change small details.
When should I pick Wan 2.7 instead?
Pick Wan 2.7 when you want a dependable, more literal result or cheaper drafts in 480p. Pick Wan 3.0 when the look matters most and you want the model to take creative risks.
How many credits does a Wan 3.0 clip use?
It depends on the clip. Longer clips, higher resolutions and native audio use more credits. The exact amount shows on the Generate button before you spend anything, and the same balance works for video, images and music.
What happens if a Wan 3.0 generation fails?
Every credit goes straight back to your balance automatically. There is nothing to claim or report. Adjust the prompt or the settings and try again.
Can I use Wan 3.0 videos commercially?
Yes. Every plan includes commercial use and downloads have no watermark. You are still responsible for what you make, so avoid real people, logos or characters you do not have the rights to.
Explore more models
One credit balance works for every model. Switch any time from the sidebar.