# Wan 3.0 Prime - Core.Today AI API

> Alibaba's Wan 3.0 Prime generates video from a text prompt (or an optional starting image) with coherent motion and lighting, at 480p, 720p or 1080p and 2 to 30 seconds long.

- **Provider**: Alibaba
- **Model ID**: alibaba/wan-3-prime
- **Category**: Video Generation
- **Credits**: 3260 per 5s 1080p video — billed per second (480p 158/s, 720p 326/s, 1080p 651/s)
- **Speed**: Slow
- **Quality**: High

## Features
- Premium tier of Wan 3.0 focused on coherent motion and lighting
- Long clips — 2 to 30 seconds in a single generation
- Three resolutions: 480p, 720p, 1080p
- Adaptive aspect ratio or one of five fixed ratios
- Negative prompts, automatic prompt expansion and seed control

## Use Cases
- Long-form (up to 30s) narrative shots without stitching
- Cinematic b-roll and establishing shots from a description
- Cost-efficient 480p drafts before a 1080p final render
- Vertical 9:16 short-form content for social media

## API Endpoint

Base URL: https://api.core.today/v1
Create Prediction: POST /predictions
Get Status: GET /predictions/{job_id}

## Authentication
Header: X-API-Key: YOUR_API_KEY

## Input Parameters

### Required
- **prompt**: string - Text prompt for video generation (up to 20,000 characters). Include scene, lighting, camera movement and action

### Optional
- **image**: string - Optional first-frame image to animate (jpg/png/bmp/webp, ≤10MB). When provided, the video is generated from this image guided by the prompt
- **resolution**: string (default: 1080p) - Output resolution. 480p bills 158 credits/sec, 720p 326/sec, 1080p 651/sec
  Options: 480p, 720p, 1080p
- **aspect_ratio**: string (default: adaptive) - 'adaptive' lets the model choose the best ratio for the prompt. Ignored when an image is provided (the image's ratio is used)
  Options: adaptive, 16:9, 9:16, 1:1, 4:3, 3:4
- **duration**: integer (default: 5) - Video duration in seconds (2-30). Billed per second of output
- **negative_prompt**: string - Content that should not appear in the video
- **enable_prompt_expansion**: boolean (default: true) - Automatically expand and optimize the prompt. Improves quality for short prompts but increases latency
- **seed**: integer - Random seed for reproducible generation (0-2147483647)

## Examples

### Text-to-Video (1080p)
A 5-second 1080p clip from a short prompt
```json
{
  "model": "alibaba/wan-3-prime",
  "input": {
    "prompt": "a man doing jumping jacks",
    "duration": 5,
    "resolution": "1080p",
    "aspect_ratio": "adaptive",
    "negative_prompt": "",
    "enable_prompt_expansion": true
  }
}
```

### Low-Cost 480p Draft
Check motion and framing at 480p before the final render
```json
{
  "model": "alibaba/wan-3-prime",
  "input": {
    "prompt": "A golden retriever running through autumn leaves in a park, camera tracking from the side, warm afternoon light",
    "negative_prompt": "blurry, distorted, low quality, static",
    "duration": 8,
    "resolution": "480p",
    "aspect_ratio": "16:9"
  }
}
```

## Response Format
```json
{
  "job_id": "abc123",
  "status": "pending | processing | completed | failed",
  "result": "URL or data (when completed)"
}
```

## Usage Flow
1. POST /predictions with model and input -> receive job_id
2. GET /predictions/{job_id} -> poll until status is completed or failed
3. Result contains output URL(s)

## Tips
- Billing is per second of output — a 30s clip costs 6x a 5s one, so draft short and extend once the motion works
- Iterate at 480p (about a quarter of the 1080p rate), then re-render the final at 1080p
- Use negative prompts such as 'blurry, distorted, low quality, static' to reduce common artifacts
- Longer durations may show motion degradation or scene drift — split very long scenes into shots
- Text rendering inside the video is unreliable; add titles in post-production

## Documentation
https://replicate.com/alibaba/wan-3-prime