AI video model by Alibaba
Wan 3.0 for smooth motion from text, photos or references
Wan 3.0 is a video model from Alibaba's Wan family. In Banana Studio it works from a written prompt, a single photo, a first and last frame, or up to three reference photos, and it can add sound to the clip.
- From 64 credits
- 5-second clips
- Up to 1080p
- Sound optional
What Wan 3.0 is good at
Every way to start
Text, one photo, a first and last frame, or up to three reference photos: Wan 3.0 supports all four creation modes in Banana Studio.
Smooth, coherent motion
Its model page describes better motion smoothness, scene fidelity and visual coherence than earlier Wan generations.
Prompt help built in
The model expands short prompts into fuller scene descriptions before it renders, so a single line can be enough to start.
Three quality levels
Draft at 480p, then render the clip you keep at 720p or 1080p.
Wan 3.0 prices in credits
| Setting | Credits |
|---|---|
| 480p | 64 |
| 720p | 128 |
| 1080p | 256 |
- Prices are per 5-second clip.
- The exact price for your settings is shown on the Create button before you start.
- The 7-day introduction ($1.99, 240 credits) covers 3 clips at the lowest setting, and the Pro plan's 2,200 monthly credits cover about 34 clips.
Plans: $1.99 for a 7-day introduction with 240 credits, then $29.99 a month for 2,200 credits, or $59.97 every 3 months with 2,200 credits each month. The introduction is a paid period, not a free trial. See plans · How credits work
Modes and shapes
| Mode | Photos | Shapes |
|---|---|---|
| Describe a video | None | 16:9, 9:16, 1:1, 4:3, 3:4 |
| Bring a photo to life | 1 | Follows your photo |
| Two photos, one video | 2 | Follows your photo |
| Put your people in a video | Up to 3 | 16:9, 9:16, 1:1, 4:3, 3:4 |
How to create with Wan 3.0
Choose how to start
Describe a video, bring a photo to life, use two photos for the first and last frame, or add up to three reference photos.
Describe the scene
One or two sentences about the subject, the action and the camera is enough; the model fills in the rest.
Pick quality and sound
Choose 480p, 720p or 1080p, decide whether to add sound, and create.
Try a prompt
Who uses Wan 3.0
Social media managers
Produce several clip ideas a day at 480p, then finish the best one in 1080p.
Brands with mascots
Keep a character or a product consistent across a series of clips with reference photos.
Teachers
Illustrate a concept or a historical scene with a short clip for a lesson or a slide.
Good to know
- Because the model expands your prompt, a very short prompt can gain details you did not ask for. Name anything that must or must not appear.
- When you animate a photo, Wan 3.0 keeps its proportions; for text prompts you can pick 16:9, 9:16, 1:1, 4:3 or 3:4.
- In the model menu it is listed as Wan 3.0.
- Wan 3.0 is made by Alibaba. Banana Studio is an independent app that offers it and is not affiliated with Alibaba.
- Page reviewed October 25, 2026.
Questions about Wan 3.0
What is Wan 3.0 good at?
Smooth, coherent motion from short prompts, with every creation mode available: text, photo, first and last frame, and reference photos. It also offers a 480p setting for cheap drafts.
How many credits does Wan 3.0 use?
Wan 3.0 costs 64 credits at 480p, 128 at 720p and 256 at 1080p per 5-second clip. Sound is optional and does not change the price. The 7-day introduction ($1.99, 240 credits) covers 3 clips at the lowest setting, and the Pro plan's 2,200 monthly credits cover about 34 clips. The exact price for your settings is shown on the Create button before you start.
Does Wan 3.0 rewrite my prompt?
It expands it. Wan 3.0 adds detail to short prompts before rendering, which helps one-line ideas look complete. If something must not appear, say so explicitly.
Can I use reference photos with Wan 3.0?
Yes. Choose Put your people in a video and add up to three photos of the people, products or characters you want to keep consistent.
Which resolution should I choose for Wan 3.0?
Use 480p to test an idea, 720p for clips watched on phones, and 1080p for larger screens. In our reviews of short clips on a phone-sized card, 1080p added little over 720p.
Can Wan 3.0 add sound to a clip?
Yes, sound is optional. Switch it on under Fine-tune your result before you create.