alibaba/wan/3.0/prime/reference-to-video
Wan 3.0 Prime by Alibaba - AI-powered video editing and transformation. Apply style transfer, motion control, lip-sync, and visual effects to existing videos with natural language instructions.
Input
Text prompt directing how the reference media is used. Reference media can be addressed positionally, e.g. 'the subject in Image 1 walks past Video 1'.
The aspect ratio of the generated image. Allowed values: 21:9, 16:9, 3:2, 4:3, 5:4, 1:1, 4:5, 3:4, 2:3, 9:16.
Include generated audio.
Output video resolution tier. Allowed values: 480p, 720p, 1080p.
230
Output duration in seconds. Set to null for smart duration, which lets the model pick a length from the prompt and reference media. Range: 2 to 30.
02147483647
Range: 0 to 2147483647.
Up to 10 reference image URLs.
Up to 5 reference audio URLs totaling at most 15 seconds.
Up to 5 reference video URLs totaling at most 15 seconds. Each clip must be at least 16 fps.
Document URL to base the video on. Requires enable_thinking=true.
Public webpage URL to base the video on. Requires enable_thinking=true. Only pages that do not require login can be read.
Enable enhanced reasoning before generation.
Idle
Example output — click Run to generate your own
Production endpoint
Send your first request
OpenAI-compatible endpoint with unified authentication and usage tracking.
POST
https://api.sandbase.ai/v1/runModel ID
alibaba/wan/3.0/prime/reference-to-video01
Input Schema
12 parameters · 1 required · 11 optional
| Parameter | Type | Required | Description |
|---|---|---|---|
prompt | string | Required | Text prompt directing how the reference media is used. Reference media can be addressed positionally, e.g. 'the subject in Image 1 walks past Video 1'. · Max length: 5000 |
seed | integer | Optional | Min: 0 · Max: 2147483647 |
audio | boolean | Optional | Include generated audio. · Default: true |
web_url | string | Optional | Public webpage URL to base the video on. Requires enable_thinking=true. Only pages that do not require login can be read. |
duration | integer | Optional | Output duration in seconds. Set to null for smart duration, which lets the model pick a length from the prompt and reference media. · Min: 2 · Max: 30 · Default: 5 |
file_url | string | Optional | Document URL to base the video on. Requires enable_thinking=true. |
resolution | string | Optional | Output video resolution tier. · Options: 480p, 720p, 1080p · Default: "1080p" 480p720p1080p |
aspect_ratio | string | Optional | The aspect ratio of the generated image. · Options: 21:9, 16:9, 3:2, 4:3, 5:4, 1:1, 4:5, 3:4, 2:3, 9:16 21:916:93:24:35:41:14:53:42:39:16 |
enable_thinking | boolean | Optional | Enable enhanced reasoning before generation. · Default: false |
reference_audio_urls | string[] | Optional | Up to 5 reference audio URLs totaling at most 15 seconds. |
reference_image_urls | string[] | Optional | Up to 10 reference image URLs. |
reference_video_urls | string[] | Optional | Up to 5 reference video URLs totaling at most 15 seconds. Each clip must be at least 16 fps. |
02
Output Schema
| Field | Type | Description |
|---|---|---|
id | string | Unique identifier for the generation task |
status | string | Task status: pending, running, completed, failed, timeout |
model | string | Model used for the generation |
outputs | array | Array of output items |
outputs[].url | string | URL of the generated artifact |
outputs[].content_type | string | MIME type (e.g. image/png, video/mp4) |
error | object | null | Error details if failed, null on success |
error.type | string | Machine-readable error type code |
error.message | string | Human-readable error description |
Async Workflow
This model uses asynchronous execution. Submit a request and poll for the result.
- Submit — POST to /v1/run, receive an
id - Poll — GET /v1/run/{id} until status is
completed,failed, ortimeout - Retrieve — Read
outputsfrom the completed response
03
Code Examples
Ready-to-run snippets
# Step 1: Submit
curl -X POST https://api.sandbase.ai/v1/run \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_API_KEY" \
-d '{
"model": "alibaba/wan/3.0/prime/reference-to-video",
"audio": true,
"duration": 5,
"resolution": "1080p",
"enable_thinking": false,
"prompt": "a beautiful sunset over mountains"
}'
# Step 2: Poll result (replace <id>)
curl https://api.sandbase.ai/v1/run/<id> \
-H "Authorization: Bearer YOUR_API_KEY"API README
Wan 3.0 Prime
Wan 3.0 Prime uses one or more subject references to produce a new sequence with the latest 3.0 family’s emphasis on coherent motion, scene detail, and cinematic presentation. Its creative focus is carrying recognizable subjects into new actions and viewpoints, with the Prime configuration favoring greater visual refinement and production polish.
Plan the clip around a clear temporal arc: define the initial state, meaningful action, camera behavior, environmental response, and intended ending. Identify each referenced subject and explain its role in the scene. Reserve Prime for material where fine detail and finishing quality justify the higher rate.
Highlights
- 3.0 subject-reference grounding. Carries referenced characters or objects into a newly staged sequence.
- Improved motion coherence. Maintains more stable subject action and environmental dynamics across the generated clip.
- Prime visual refinement. Prioritizes fine textures, lighting, composition, and cleaner production detail.
- Viewpoint-flexible identity. Keeps recognizable traits as referenced subjects change pose, action, and camera angle.
Pricing
| Configuration | Billing unit | Price |
|---|---|---|
| Generated duration | 480p, per second | $0.075 |
| Generated duration | 720p, per second | $0.150 |
| Generated duration | 1080p, per second | $0.300 |
When to Use
| ✅ Good fit | ❌ Consider alternatives |
|---|---|
| The project needs this exact reference-to-video workflow | The intended task belongs to a different media route |
| Available source media matches every required field | Required assets or usage rights are unavailable |
| The brief can define subject, action, camera, and style | Output must be deterministic at frame or pixel level |
| Supported duration, resolution, and framing fit delivery | Final placement requires unsupported specifications |
| An asynchronous generation job fits production | A live or frame-synchronous response is mandatory |
Prompt Guide
Describe the result as a shot or design brief: identify subjects and references, state the action or transformation, specify environment and composition, then add camera behavior, lighting, pacing, style, sound, and preservation constraints where relevant. Use exact reference identifiers exposed by the local schema.
{
"aspect_ratio": "21:9",
"audio": true,
"prompt": "A cinematic scene with clearly directed subject action, camera movement, lighting, pacing, and atmosphere",
"resolution": "480p"
}
Technical Specs
| Spec | Value |
|---|---|
| Model ID | alibaba/wan/3.0/prime/reference-to-video |
| Input fields | prompt (string)<br>aspect_ratio (string; 21:9, 16:9, 3:2, 4:3, 5:4, 1:1, 4:5, 3:4, 2:3, 9:16)<br>audio (boolean)<br>resolution (string; 480p, 720p, 1080p)<br>duration (integer; 2–30)<br>seed (integer; 0–2147483647)<br>reference_image_urls (array)<br>reference_audio_urls (array)<br>reference_video_urls (array)<br>file_url (string)<br>web_url (string)<br>enable_thinking (boolean) |
| Required input | prompt |
| Output fields | url, content_type |
| Execution | Asynchronous job |
Related Models
alibaba/wan/3.0/prime/image-to-video— Compare this concrete local family route.alibaba/wan/3.0/prime/text-to-video— Compare this concrete local family route.alibaba/wan/3.0/image-to-video— Compare this concrete local family route.