# Wan 3.0 for Product Videos: What's Live and What Isn't Yet

> For software teams tracking new video models: what Wan 3.0 actually ships today, what it costs, and which parts still aren't wired into its public integrations.

**Key takeaways**

- Wan 3.0 went live on ImagineToVideo on October 21, 2026, offering text/image-to-video up to 1080p and 30 seconds.
- Wan 3.0 costs $0.05-$0.20 per second on fal depending on resolution, with five free 720p generations daily.
- Wan 3.0's reference-to-video mode, needed to anchor clips to real screens, isn't wired into ImagineToVideo yet.

Published Oct 25, 2026 · uscale · https://uscale.ai/blog/wan-3-0-product-videos

## Wan 3.0 for Product Videos: Live, but Not Finished

Wan 3.0 for product videos went live on ImagineToVideo on October 21, 2026, and the headline spec is generous: text-to-video and image-to-video clips up to 1080p and up to 30 seconds, with sound.

Then comes the line further down the [Wan 3.0 launch notes on ImagineToVideo](https://imaginetovideo.com/blog/wan-3-0-now-live-specs-pricing-how-to-use). Wan 3.0's API has a separate reference-to-video mode, and it isn't wired into this integration yet. That mode is the part that would let you hand the model your real screens and keep them steady across a generated clip.

So the model is live everywhere except the part a product video needs most. Whether that gap matters depends on which shots you were hoping it would make.

Here's the short answer first, then the specs and prices behind it.

[Video: Cinematic for a GPU marketplace app](https://uscale.ai/examples/compute-exchange-cinematic.mp4) — A narrated product film made from the real product — not a Wan 3.0 clip, which is what this post is weighing it against.

## What Wan 3.0 Means for Product Videos

> **Wan 3.0 is cheap, long and sharp enough for the motion around a product video, but not for the product itself: the reference-to-video and Prime modes that would anchor a scene to your real screens aren't in the public integration this launch covers.** So skip it for UI capture and consider it for background motion: mood shots, transitions, social loops.

That verdict rests on what the model can make today, so start there.

## What Can Wan 3.0 Actually Make Right Now?

Right now Wan 3.0 makes text-to-video and image-to-video clips up to 1080p and up to 30 seconds, with sound, in the [ImagineToVideo generator](https://imaginetovideo.com/blog/wan-3-0-now-live-specs-pricing-how-to-use). You can also add a second image as an end frame. The clip then animates from the first image into the second, which suits a transition or a before-and-after shot.

![Wan 3.0 on ImagineToVideo at launch: Max resolution 1080p, Max length with sound 30 seconds](https://uscale.ai/blog-media/wan-3-0-product-videos/chart-wan-3-0-on-imaginetovideo-at-launch-1200x396.svg "Source: ImagineToVideo / Wan 3.0 Is Live on ImagineToVideo: Specs, Pricing and How to Use It, Oct 21, 2026")
### The settings you can change today

- **Resolution:** 480p, 720p (the default) or 1080p.
- **Length:** 3 to 15 seconds, or 20 or 30 seconds. The default is 5.
- **Aspect ratio, for text-to-video:** 16:9, 9:16, 1:1, 4:3 or 3:4.

Most people test at the defaults and then plan their shot list around what they saw. That gets it backwards. A short default clip tells you whether a prompt works, but it doesn't tell you how much the model can carry in one take. Plan against the ceiling instead. One pass at 1080p and 30 seconds is the longest, sharpest clip you can ask for. Anything longer has to be an edit of several clips, not a single generation.

That ceiling only helps if the per-second price lets you use it. Price is the next question.

## How Much Does Wan 3.0 Cost?

Wan 3.0 costs $0.05 to $0.20 per second of output on fal, depending on resolution: $0.05 at 480p, $0.10 at 720p and $0.20 at 1080p. A 5-second 720p clip runs $0.50, and a 10-second 1080p clip runs $2.00. fal notes the rates can change.

![Wan 3.0 pricing by resolution (fal's public API): 480p $0.05/s, 720p $0.10/s, 1080p $0.20/s](https://uscale.ai/blog-media/wan-3-0-product-videos/chart-wan-3-0-pricing-by-resolution-fal-s-publ-1200x400.svg "Source: fal.ai / Wan 3 - Alibaba's Next-Generation AI Video Model, Oct 1, 2026")
You can test before you pay. A signed-in fal account gets five free Wan 3 text-to-video generations every 24 hours in the fal sandbox, at 720p with audio, on a rolling window. That is enough to find out whether a prompt works before it costs anything.

If you would rather not use an API, ImagineToVideo lists Wan 3.0 as the cheapest model on its site:

- 20 credits for a 5-second 720p clip
- plans from 10 credits for a 5-second 480p clip
- a $0.99 Starter Pack that covers one 5-second 720p clip

### How to budget a month of tests

Work out the prompt for free, then pay for resolution once. Spend the daily free generations on wording and framing. Move a prompt to 1080p only when it is the one you will use, and treat each 10-second 1080p render as a $2.00 decision.

**The per-second rate is low. The waste comes from paying 1080p rates to find out that a prompt doesn't work.**

What the low price doesn't buy is the next question.

## Is Wan 3.0 Ready to Replace a Filmed Product Demo?

Not yet. Its strength is real: up to 30 seconds at 1080p with sound, at the cheapest rate of any model on ImagineToVideo. But that strength is generated motion. Wan 3.0 makes new footage. It does not capture your product, and the modes that would anchor a clip to your real screens sit outside this launch.

fal's reference-to-video mode takes up to 10 reference images, up to 5 video clips totalling 15 seconds at 16fps or higher, and up to 5 audio tracks totalling 15 seconds. ImagineToVideo hasn't wired it in. QwenCloud's release log shows the Prime variant only as an invite-only test entry (August 6, 2026) and a general-availability entry (August 24, 2026), with no licence or parameter count stated.

| What you need | Wan 3.0 on ImagineToVideo today |
|---|---|
| Text- or image-to-video, up to 1080p and 30 seconds | Live |
| Reference-to-video (anchor a clip to real screens) | Not wired in |
| Prime reference-to-video | Separate on fal: $0.068/s at 480p, $0.14 at 720p, $0.28 at 1080p |
| 21:9 framing | Not available yet |
| 2-second clips | Not available yet |

Cheap, long generated motion is what makes Wan 3.0 right for the b-roll around a demo. It is also why it can't be the demo.

So what do you actually do with that on your next video?

## What This Means for Your Next Product Video

Split your next video into two jobs before you open any tool: the motion around the product, and the product itself.

**Use Wan 3.0 for the motion around the product.** That covers a mood shot to open, a transition between scenes or a background loop under a headline. The end-frame option animates one image into a second, which suits a before-and-after shot. Quick win for this hour: run one b-roll prompt in fal's sandbox, which gives signed-in accounts five free 720p generations with audio every 24 hours. You're done when one clip is good enough to cut in.

**Keep the product's own screens out of the generator.** Reference-to-video isn't wired into the integration this launch covers, and a generated interface can drift from the one you ship.

**Pick how you film the screens.** The free route is the screen recorder you already have and a script in a doc. The other route is a tool that films the product rather than generating it. uscale opens your live site or signed-in app, films each screen and returns a narrated, captioned video once you approve the storyboard in a chat. uscale says this takes minutes. [Software teams](https://uscale.ai/industries/software-products) can start from the [AI product video generator](https://uscale.ai/use-cases/ai-product-video-generator) page, and the [example films](https://uscale.ai/examples) show real results. You're done when every must-show shot is filmed.

![Filming the product instead of generating it: 1. Paste your site, app or GitHub repo 2. Approve the storyboard in the chat 3. Get a narrated film of your real screens](https://uscale.ai/blog-media/wan-3-0-product-videos/steps-filming-the-product-instead-of-generatin-1200x360.svg)
If you do only one thing, list the shots that must show your real product. Success looks like real screens in the walkthrough and generated motion only around them.

**Try it:** [make an AI product video](https://uscale.ai/use-cases/ai-product-video-generator) of your real screens, and approve the storyboard before anything is filmed.

## The Bottom Line on Wan 3.0 for Product Videos

Three things to do this week:

1. **Use Wan 3.0 for the motion around your product.** A mood shot, a transition, a background loop. Keep it away from the product's own screens until reference-to-video reaches the integration you use.
2. **Film the real screens with a tool built to film them.** uscale opens your live site or app, films each screen and adds narration and captions once you approve the storyboard. Your first video is free. You can [see the free tier on the pricing page](https://uscale.ai/pricing) before you commit to either route.
3. **Bookmark this page.** It gets updated if Wan 3.0 joins uscale's film styles.

A generated clip can set the mood. Only the product itself can show what it does.

## Questions

### What is Wan 3.0?

Wan 3.0 is a video generation model that turns a text prompt or an image into a clip up to 1080p and up to 30 seconds long, with sound. It went live on ImagineToVideo on October 21, 2026, and fal offers public text-to-video and image-to-video endpoints for it. An optional end frame lets a clip animate from one image into a second, which suits transitions and before-and-after shots.

### How much does Wan 3.0 cost per video?

Wan 3.0 is billed per second of output on fal: $0.05 at 480p, $0.10 at 720p and $0.20 at 1080p, and fal notes the rates can change. A 5-second 720p clip costs $0.50 and a 10-second 1080p clip costs $2.00. On ImagineToVideo, a 5-second 720p clip costs 20 credits, and a $0.99 Starter Pack covers one such clip. Signed-in fal accounts also get five free 720p text-to-video generations every 24 hours.

### Is Wan 3.0's reference-to-video mode available yet?

Wan 3.0's reference-to-video mode exists in its API, and fal lists it as accepting up to 10 reference images, up to 5 video clips totalling 15 seconds and up to 5 audio tracks totalling 15 seconds. ImagineToVideo has not wired it into its integration yet. A separate Prime reference-to-video variant is billed on fal at $0.068 per second at 480p, $0.14 at 720p and $0.28 at 1080p.

### Can Wan 3.0 film my actual product instead of generating a scene?

No. Wan 3.0 generates new footage from a prompt or an image; it does not capture a working product. The reference-to-video mode that would anchor a clip to real screens isn't wired into ImagineToVideo's integration yet, and a generated interface can drift from the one you ship. To show a real product, record its screens yourself or use a tool that films the live site or app, such as uscale.

### How much does a product video cost with uscale instead?

A uscale product video starts free: the first video is free, with 10 credits on the free plan, or 20 with a work email, and no card needed. After that, credits are bought once with no subscription: Lite is $29 for 25 credits, about 2 product walkthroughs; Starter is $49 for 50 credits, about 5; Growth is $129 for 150 credits, about 15. The director names each price before it starts.

---
Source: https://uscale.ai/blog/wan-3-0-product-videos