Wan 3.0: Specs, Pricing and What Is Actually Confirmed
Alibaba's Wan 3.0 runs on ImagineToVideo since October 21, 2026: text to video and image to video with an optional end frame, 480p to 1080p, 3 to 30 seconds, with native audio.
Below: what you can generate with Wan 3.0 here and what it costs, then a sourced fact-check of the specs circulating online. Every line names its source, and every source has a check date. Last verified October 21, 2026.
Wan 3.0 is live in our generator
Since October 21, 2026 you can generate with Wan 3.0 on ImagineToVideo, through fal's public API: text to video, or image to video from a start frame with an optional end frame, at 480p, 720p or 1080p, 3 to 15, 20 or 30 seconds, audio on by default. A 5-second 720p clip costs 20 credits. The fact-check further down covers Alibaba's own access and documentation separately.
All sources last checked on
Official video
Alibaba's own Wan 3.0 announcement footage
The launch video Alibaba published for Wan 3.0 on August 6, 2026. Every shot in it was rendered by Alibaba, not on ImagineToVideo. Treat it as vendor marketing, not as a benchmark: the generator is the place to judge the model on your own prompts.
Wan 3.0 on ImagineToVideo: what you can set
Text or image to video
Describe a scene, or upload a start frame. Add an optional end frame and Wan 3.0 animates from one to the other. Image-to-video follows the shape of your source image.
3 to 30 seconds in one pass
Pick any length from 3 to 15 seconds, or 20 or 30 seconds. Each clip is a single generation, not stitched segments.
480p, 720p or 1080p
720p is the default. Aspect ratios for text-to-video are 16:9, 9:16, 1:1, 4:3 and 3:4. 4K and 21:9 are not offered here.
Audio included, lowest price on the site
Sound is generated with the video by default. A 5-second clip costs 10 credits at 480p, 20 at 720p and 40 at 1080p, which makes Wan 3.0 the cheapest model in our generator.
Confirmed, with the source for each line
Every statement below comes from Alibaba, QwenCloud or fal directly. Nothing here is sourced from an aggregator or a third-party spec page.
Wan 3.0 launched on QwenCloud on August 6, 2026 as an invite-only test, and became generally available on August 24, 2026.
QwenCloud's model release log carries both entries. The August 24 entry reads "Initially launched as invite-only on August 6, now publicly available", and adds a faster variant, wan3.0-video-prime.
The documented model identifier is wan3.0-video.
A create-task API reference page exists under that identifier, which is why the parameters below can be quoted rather than inferred.
Clip length runs from 2 to 30 seconds, with a smart-duration option.
The duration parameter accepts an integer in that range. Supplying a source video does not extend the ceiling — the combined output still stops at 30 seconds.
The documented resolution options are 480P, 720P and 1080P, defaulting to 1080P.
This matters because it is the single most misreported figure about this model. See the claims table below for what that does and does not settle.
Seven reference input fields are documented, including files and links.
First frame, last frame, reference image, reference video, reference audio, file and link. Aspect ratio can be set to adaptive or one of six fixed ratios, including 21:9, and the prompt field accepts up to 20,000 characters. ImagineToVideo currently wires the first and last frame only.
fal publishes Wan 3.0 at $0.05 per second at 480p, $0.10 at 720p and $0.20 at 1080p.
Those are the rates ImagineToVideo pays and converts into credits. QwenCloud's own API reference still states no price. ⚠️ Per-second figures quoted for direct QwenCloud access come from third-party pages and are not confirmed here.
Claims in circulation, checked one by one
These are the Wan 3.0 specs being repeated across the web. For each one, what the official sources actually say.
| Claim in circulation | What the official sources say | Verdict |
|---|---|---|
Native 4K output | Contradicted by the documented parameter list This settles what the API accepts today. It is not a statement about what the model may support later. | |
Released under Apache 2.0 as open weights | ⚠️ Unconfirmed — nothing published so far Absence to date. It does not rule out a later release. | |
27B or 60B mixture-of-experts architecture | ⚠️ Unconfirmed — no official figure exists | |
An "Identity Lock" character-consistency feature | ⚠️ Unconfirmed — name not used in any official source | |
An "AI Director" shot-planning mode | ⚠️ Unconfirmed — name not used in any official source | |
Open to the public | Confirmed since August 24, 2026 Invite-only applied only to the first two and a half weeks. |
Every verdict above reflects the state of the official sources on 2026-10-21.
The sources, and what each one showed
The places a Wan 3.0 claim can be settled, and what was found in each on the verification date.
- Checked
- 2026-10-21
- What it showed
- Entry dated 2026-08-06 (invite-only test) and entry dated 2026-08-24 (general availability, plus wan3.0-video-prime). No pricing, licence or parameter count stated.
- Checked
- 2026-10-21
- What it showed
- Live, with a full parameter list: 2–30s duration, 480P/720P/1080P resolution, adaptive plus six fixed aspect ratios, seven reference inputs, audio on by default. No pricing.
- Address
- Checked
- 2026-10-21
- What it showed
- No Wan 3.0 weights. The newest additions are Wan2.2-Animate-2 models, published in August 2026.
- Checked
- 2026-10-21
- What it showed
- No Wan 3.0 repository. The most recently updated repository is Wan2.2, last touched 2026-09-21.
- Checked
- 2026-10-21
- What it showed
- Public text-to-video and image-to-video endpoints. $0.05/s at 480p, $0.10/s at 720p, $0.20/s at 1080p, subject to change.
Anything that changes in these sources is written into the update log at the bottom of this page on the same day.
What this means for you
If you make videos
Open the generator, choose Wan 3.0 and write a prompt or upload a start frame. A 5-second 720p clip costs 20 credits, so the $0.99 Starter Pack covers your first one. Judge the model on your own prompts rather than on Alibaba's launch footage.
If you build on video APIs
Wan 3.0 is generally available on QwenCloud and on fal, with a documented request schema. Two things are still worth watching: whether the resolution list grows beyond 1080P, and whether weights ever appear on Hugging Face or GitHub, since that decides whether self-hosting is possible.
How to generate with Wan 3.0
From prompt or start frame to a clip with audio in four steps
Open the Wan 3.0 generator
Go to text to video or image to video and pick Wan 3.0 from the model list. It carries the Cheapest badge.
Write a prompt or add frames
For text to video, describe the scene, camera and sound. For image to video, upload a start frame and, if you want a set ending, an end frame.
Set length and resolution
Choose 3 to 15, 20 or 30 seconds and 480p, 720p or 1080p. The credit cost updates as you change them.
Generate and download
Start the job. Clips arrive with audio and can be downloaded or reused from your library. Failed jobs refund their credits.
Update log
Every change to the official sources, or to how Wan 3.0 runs on ImagineToVideo, lands here with its date.
Wan 3.0 goes live on ImagineToVideo
Text-to-video and image-to-video with an optional end frame, 480p to 1080p, 3 to 30 seconds, through fal's public API. Re-check of all sources: QwenCloud's log shows general availability since August 24; still no weights and no 4K value in the API.
Wan 3.0 becomes generally available on QwenCloud
The release log moves wan3.0-video from invite-only to publicly available and adds wan3.0-video-prime, a faster variant.
Page published after a full check of all four sources
The API reference was found to be live with a complete parameter list, which moved the 4K claim from unverified to contradicted. Open weights remain absent from both Hugging Face and GitHub. No pricing published.
wan3.0-video appears in QwenCloud's release log
Listed as an invite-only test covering text-to-video, image-to-video from a first frame or a first-and-last frame pair, and reference-to-video in one unified model.
Questions people are asking about Wan 3.0
When was Wan 3.0 released?
QwenCloud's model release log dates wan3.0-video to August 6, 2026, as an invite-only test, and marks it generally available on August 24, 2026. ImagineToVideo added it on October 21, 2026. Anything describing a release before August 6 refers to a preview, an ecosystem model such as Wan-Dancer, or speculation.
Can I use Wan 3.0 right now?
Yes. Wan 3.0 is in the ImagineToVideo generator for text to video and image to video, with an optional end frame. Pick it from the model list on either page. A 5-second 720p clip costs 20 credits, and the $0.99 Starter Pack covers it.
Is Wan 3.0 open source?
No weights have been published as of October 21, 2026. There is no Wan 3.0 repository on the Wan-AI organisation on Hugging Face or the Wan-Video organisation on GitHub. Earlier Wan models were released openly, which is probably why the Apache 2.0 claim spread, but it has not happened for this version.
Does Wan 3.0 support 4K?
Not through the documented API. QwenCloud's create-task reference enumerates resolution as 480P, 720P and 1080P, with 1080P as the default — 4K is not an accepted value. That settles what the API offers today; it is not a statement about what the model might support in a later revision.
How much does Wan 3.0 cost?
On fal, Wan 3.0 costs $0.05 per second at 480p, $0.10 at 720p and $0.20 at 1080p. On ImagineToVideo that becomes credits: a 5-second clip is 10 credits at 480p, 20 at 720p and 40 at 1080p, and a 30-second 1080p clip is 270. ⚠️ QwenCloud's own API reference states no price.
Which Wan 3.0 features are not on ImagineToVideo yet?
Reference-to-video, which takes reference images, video and audio, is not wired in yet. The 21:9 aspect ratio and 2-second clips are also not offered here. Text to video, image to video and first-to-last-frame generation are all available.
Generate your first Wan 3.0 video
Up to 30 seconds with sound, at up to 1080p, for less than any other model on the site. Start with the $0.99 Starter Pack, no subscription needed.