Mixer AI
Wan 3.0
A scene up to 30 seconds long, combining photos, motion and audio references in one request.
When to choose it
Try Wan 3.0 when a brief animated image is not enough and the action needs a beginning and an ending. Anchor first and last frames or build a scene from several reference types. Plan longer clips as a sequence: who does what, and where the action ends.
- Use different reference types for appearance, movement and sound.
- Up to 30 seconds leaves room for one event to develop beyond a single gesture.
Modes and settings
- Text, one or two frames, or references. Mixer AI offers Standard and Prime.
- 480P, 720P or 1080P; reference mode accepts up to 10 photos, 5 videos and 5 audio files.
- Each video or audio file is 1–15 seconds, with a 15-second total limit for each type. Audio requires a photo or video.
Price in coins
| Version | Quality / size | Input | Coins / second |
|---|---|---|---|
| Standard | 480P | With video reference | 5.16 |
| Standard | 480P | Text or images | 5.16 |
| Standard | 720P | With video reference | 11.61 |
| Standard | 720P | Text or images | 11.61 |
| Standard | 1080P | With video reference | 20.64 |
| Standard | 1080P | Text or images | 20.64 |
| Prime | 480P | With video reference | ≈ 7.89 |
| Prime | 480P | Text or images | ≈ 7.89 |
| Prime | 720P | With video reference | ≈ 16.25 |
| Prime | 720P | Text or images | ≈ 16.25 |
| Prime | 1080P | With video reference | ≈ 32.51 |
| Prime | 1080P | Text or images | ≈ 32.51 |
Billing is per second, not per fixed-length clip. The rate is multiplied by the billable duration; the complete request total is rounded up to a whole coin once.
Without video input, billing uses the output duration. With a video reference, it uses input video time plus output time at the video-reference rate. Photos and audio do not add video seconds.
≈ marks a fractional rate rounded to two decimal places for display; billing uses full precision.
What you can make
- A product demonstration with several stages.
- A scene with a planned opening, movement and ending.
How to direct the result
- Break a longer scene into consecutive events and assign every reference a role.
- State what must stay: identity, clothing, product shape and movement direction.
Example: a desk lamp demonstration
10 seconds. Use the lamp photo for shape and material. 0–3s: a wide shot of the desk. 3–7s: a hand presses the switch and warm light comes on. 7–10s: the camera moves toward the lit page. Keep the lamp's construction unchanged. Switch click and quiet room tone.
What to keep in mind
- Input video time is added to output time in the price calculation.
- A long generation is not precise editing. Inspect transitions and object consistency throughout.
FAQ
Can I use an audio reference alone?
No. Add an image or video alongside it.
Can all five videos be fifteen seconds long?
No. Their combined duration is also limited to fifteen seconds.
Can I specify the last frame?
Yes, in Frame mode. Prepare opening and ending images with compatible compositions.
How do Standard and Prime differ?
They are separate model variants with different tariffs. Compare them on one short scene; a higher price does not guarantee success.
Why does input video increase the total?
With a video reference, billing counts input video time plus output time. The table shows the tariff per billable second without assuming a clip length.
Must I generate thirty seconds?
No. Choose the time the action needs. A short test is easier to evaluate.
What changes the price?
The coin rate per second depends on quality, mode and, for some models, sound. Multiply it by the billable duration. Where input video is charged, its duration also counts. The complete request total is rounded once.
Do I need a separate model subscription?
No. In Mixer AI, you top up one coin balance and use it for the tasks you need. You do not need a separate subscription to this model.