Wan 3.0

Wan 3.0 (alibaba/wan-3.0) is a video generation model from Alibaba for text-to-video, frame-to-video, and reference-guided generation. It produces 480p, 720p, or 1080p video with durations from 1 to 15 seconds. The fast mode with the same geometry is Wan 3.0 Prime (alibaba/wan-3.0-prime): the same model delivered faster. Generation is asynchronous: POST /v1/videos returns a job, then you poll the status and download the mp4.

Parameters

Field	Description
model	alibaba/wan-3.0 (or the short wan-3.0) or alibaba/wan-3.0-prime
prompt	Scene description, required; up to 5000 characters
seconds	Any integer from 1 to 15; defaults to 5
size	A WxH size from the table below; defaults to 1280x720
input_reference	Image-to-video: one image as the starting frame, {"image_url": "data:image/...;base64,..."}
input_references	Reference-to-video: 1 to 10 images in an array, also data: URLs only
input_video	Video reference: {"video_url": "https://..."} or a data: URL — motion is carried over to the character in your images; up to 3 clips in the dashboard and bot, one in the API
input_audio	Audio reference: {"audio_url": "https://..."} or a data: URL, up to 3 tracks — free of charge
seed	Reproducibility seed: an integer 0…2147483647. The same prompt with the same seed produces the same clip; without it the upstream picks a random one

Six allowed sizes — two aspect ratios in three resolutions:

Aspect	480p	720p	1080p
16:9	852x480	1280x720	1920x1080
9:16	480x852	720x1280	1080x1920

A value outside the list returns 400 with the allowed options — check the sizes field in the catalog if in doubt.

Examples

Text-to-video

Python

import requests

job = requests.post(
    "https://api.mixen.ai/v1/videos",
    headers={"Authorization": f"Bearer {API_KEY}"},
    json={
        "model": "alibaba/wan-3.0",
        "prompt": "A metro train emerges from a tunnel onto a bridge as the city wakes up, long camera glide",
        "seconds": "12",
        "size": "1280x720",
    },
).json()
print(job["id"], job["status"])

cURL

curl https://api.mixen.ai/v1/videos \
  -H "Authorization: Bearer $MIXEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "alibaba/wan-3.0",
    "prompt": "A metro train emerges from a tunnel onto a bridge as the city wakes up, long camera glide",
    "seconds": "12",
    "size": "1280x720"
  }'

Frames: first and last

The clip starts on the given frame and ends on another given image — the end frame is available in the dashboard and bot (a second photo); the API does not have a separate field for it yet:

import base64
import requests

with open("start.jpg", "rb") as f:
    img = "data:image/jpeg;base64," + base64.b64encode(f.read()).decode()

job = requests.post(
    "https://api.mixen.ai/v1/videos",
    headers={"Authorization": f"Bearer {API_KEY}"},
    json={
        "model": "alibaba/wan-3.0",
        "prompt": "The camera glides along a street from dawn to evening lights",
        "seconds": "8",
        "size": "1280x720",
        "input_reference": {"image_url": img},
    },
).json()
print(job["id"], job["status"])

Reference-to-video with several images

From one to ten images — the model assembles the scene while keeping the characters and style of the references:

{
  "model": "alibaba/wan-3.0",
  "prompt": "The hero from the first frame walks through the alley from the second one, neon signs",
  "seconds": "6",
  "size": "1280x720",
  "input_references": [
    "data:image/png;base64,iVBORw0...",
    "data:image/png;base64,iVBORw0..."
  ]
}

Then as usual: GET /v1/videos/{id} every 5–10 seconds until completed, then GET /v1/videos/{id}/content for the mp4. For the full cycle see the video guide.

Pricing

Billed per second, the rate depends on the resolution. Wan 3.0: 480p — 5.43 ₽ per second, 720p — 10.86 ₽, 1080p — 21.73 ₽. Wan 3.0 Prime: 480p — 7.37 ₽, 720p — 15.19 ₽, 1080p — 30.46 ₽. Audio references are free; a video reference is billed the same per-second rate for the clip’s own seconds. Current prices — the catalog.

Strengths and limitations

• A continuous 1–15 second duration — both short cuts and medium shots with one model.
• Two frame formats — landscape and vertical, three resolutions up to 1080p.
• The first → last frame transition is supported (dashboard and bot).
• Three kinds of references: up to 10 images, up to 3 clips and up to 3 audio tracks (audio is free).
• Prime is the fast mode: the same parameters at a higher rate when the clip is needed sooner.
• The 1:1, 4:3 and 3:4 formats are not supported; neither are clips longer than 15 seconds.

All models — in the catalog. The general video workflow — the video guide.