Wan 3.0

Node type: video:wan_v3_0_t2v
Category: video

Description

Alibaba Wan 3.0: video with sound from text, a start frame (plus an optional end frame), or reference images and audio. 2 to 30 seconds at 480p, 720p or 1080p.

Pricing

• Cost: ~12 credits per seconds
• Per-second rate: The figure is the rate for a 5-second clip. Shorter clips can cost a bit more per second. You're always charged for the real length.

Variant family

• Family: wan_v3_0
• Is primary: yes
• All variants: video:wan_v3_0_t2v, video:wan_v3_0_i2v, video:wan_v3_0_r2v

Canvas ports

These appear as port handles on the left side of the node.

ID	Label	Details
prompt	Prompt	TEXT (required) · multiline
image_url	Start Frame	IMAGE
end_image_url	End Frame	IMAGE
reference_image_urls	Reference Images	IMAGE · list
reference_audio_urls	Reference Audio	AUDIO · list

Sidebar config

These render as form fields in the right-side config panel when the node is selected.

ID	Label	Details
aspect_ratio	Aspect Ratio	TEXT · options: adaptive, 16:9, 9:16, 1:1, 4:3, +1 — Adaptive lets the model pick the shape, or follow the start frame when one is wired.
duration	Duration (seconds)	NUMBER · 2…30
resolution	Resolution	TEXT · options: 480p, 720p, 1080p
audio	Generate Audio	BOOLEAN — Sound is generated with the video. Same price either way.
enable_prompt_expansion	Prompt Expansion	BOOLEAN — Rewrites the prompt before generating. Turning it off is 20 to 60 seconds faster but usually looks worse.
enable_thinking	Thinking	BOOLEAN — Plans the shot before generating. Slower.
seed	Seed	NUMBER · 0…2147483647

Outputs

ID	Label	Type
video	Generated Video	VIDEO

One node, three endpoints

Alibaba ships Wan 3.0 as three APIs. This is one node, and what you wire picks the API:

• just a Prompt goes to text-to-video
• a Start Frame goes to image-to-video, and that image becomes the first frame. Add an End Frame to set the last one too
• anything in Reference Images goes to reference-to-video, with optional Reference Audio beside it

References win over a start frame. Wire both and the run goes to reference-to-video, with the start frame added to the end of the reference list instead of being dropped.

Two wirings are refused before you are charged, with a message that says what to add:

• Reference Audio with no Reference Image and no Start Frame. The audio needs at least one picture next to it.
• an End Frame with no Start Frame. The end frame only works as the last frame of an image-to-video run.

Wan 3.0 or Wan 3.0 Prime

Two nodes with the same ports and settings. Prime is the higher quality tier and costs about 40% more. To switch, add the other node and wire it the same way.

What a run costs

Billed per second of video, and a higher resolution costs more per second. A 5 second clip at 720p is 58 credits on Wan 3.0 and 81 on Prime. Sound costs nothing extra. The card shows the exact price before you run.

References

Name them in the prompt by position: "Image 1", "Audio 1". Up to 10 images and 5 audio clips (15 seconds of audio combined) reach the model. Anything past that is trimmed from the end of the list.

No reference video yet

The API also takes reference video clips, and this node does not offer them yet. Alibaba bills the length of a reference clip on top of the video it renders, and nothing measures a clip before you are charged, so the price would be wrong. Reference images and audio are not affected.

Duration

2 to 30 seconds, in whole seconds. A fraction (say 2.6, set through the API or an agent) is refused before you are charged rather than rounded, so the price you saw is the price you pay. The model can also pick its own length, but a length nobody knows up front cannot be priced, so the node always asks for the duration you set.

---

Auto-generated from the Wireflow node registry.