# Alibaba Wan 3.0 Pro Video API

> Alibaba Wan 3.0 Pro Video generates videos up to 30 seconds with native audio from text, a first/last frame, or reference images, videos, and audio, upscaled to 1080p, 2K, or 4K.

Source: https://www.sandbase.ai/models/alibaba/wan/3.0/pro/video
API reference: https://www.sandbase.ai/models/alibaba/wan/3.0/pro/video#api

## Facts

- Provider: Alibaba
- Type: Video model
- Model ID: `alibaba/wan/3.0/pro/video`
- Price: From $0.90 per run
- Capabilities: text to video, image to video, reference to video, video to video, audio to video
- Billing: pay per use from one SandBase balance; no provider account or subscription needed

## Call it

- Unified Run API: `POST https://api.sandbase.ai/v1/run`

```bash
curl -X POST https://api.sandbase.ai/v1/run \
  -H "Authorization: Bearer $SANDBASE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "alibaba/wan/3.0/pro/video",
  "image": "https://static.sandbase.ai/samples/image-to-video/i2v-black-cat-sky-img0.png",
  "prompt": "A cinematic tracking shot of a sailboat crossing a glowing sea at sunset",
  "duration": 5,
  "resolution": "1080p",
  "generate_audio": true
}'
```

The response returns a run id; poll `GET https://api.sandbase.ai/v1/run/{id}` until the status is `completed` or `failed`.

## FAQ

### How do I call Alibaba Wan 3.0 Pro Video through SandBase?

Create a SandBase API key, then send requests with the model ID "alibaba/wan/3.0/pro/video" to Unified Run API (/v1/run) at https://api.sandbase.ai. The request examples on this page show the exact payload.

### Do I need a separate Alibaba account?

No. One SandBase API key and balance gives you access to Alibaba Wan 3.0 Pro Video and the other models in the catalog; you do not need to sign up with Alibaba separately.

## README

### Wan 3.0 Pro Video

A single endpoint for the whole Wan 3.0 Pro model. Send only a prompt for text-to-video, a first/last frame for image-to-video, or any mix of reference images, videos, and audio for reference-driven generation. Clips run 2–30 seconds at 30fps with sound generated alongside the picture. Pro runs the same Wan 3.0 model, then upscales the result with FlashVSR, so output comes back at 1080p, 2K, or 4K. For 480p or 720p use the non-Pro endpoints.

#### Pick an input mode

| Mode | Send | Can't combine with |
|---|---|---|
| Text | `prompt` only | — |
| Frames | `image`, optional `end_image` | `images`, `videos`, `audios` |
| References | any of `images`, `videos`, `audios` | `image`, `end_image` |

Mixing frames and references is rejected with a 400 before anything is billed. A reference video plus an edit or continuation instruction also works here; the dedicated video-edit and video-extend endpoints give those cases a simpler schema.

#### Fields

| Field | Type | Default | Notes |
|---|---|---|---|
| `prompt` | string, required | — | Up to 5,000 characters. |
| `image` | image URL | — | First frame. JPEG/PNG/BMP/WebP, 240–8000px per side, ≤20MB, no transparency. |
| `end_image` | image URL | — | Last frame; needs `image`. |
| `images` | image URL[] (≤10) | — | Referenced as Image 1, Image 2, … |
| `videos` | video URL[] (≤5) | — | MP4/MOV, 1–15s each, 15s total, ≤100MB. Referenced as Video 1, … |
| `audios` | audio URL[] (≤5) | — | WAV/MP3, 1–15s each, 15s total, ≤15MB. Referenced as Audio 1, … |
| `duration` | 2–30 or `auto` | 5 | `auto` lets the model pick the length. |
| `resolution` | `1080p` / `2k` / `4k` | `1080p` | Output tier after FlashVSR super-resolution; sets the per-second rate. |
| `aspect_ratio` | `adaptive` / `16:9` / `4:3` / `1:1` / `3:4` / `9:16` | `adaptive` | `adaptive` follows the prompt and inputs. |
| `generate_audio` | boolean | true | Native dialogue, music, and effects. Same price on or off. |
| `seed` | integer 0–2147483647 | random | Fix it to compare prompt changes. |

#### Billing

Following the Wan 3.0 Pro rule, billed seconds are the output length plus the total length of any reference `videos` (rounded up, counted up to 15s): $0.18 at 1080p, $0.20 at 2K, $0.23 at 4K. Input images and audio are not metered. If the reference video length cannot be read, 15s is assumed.

| Request | Billed | 1080p | 2K | 4K |
|---|---|---|---|---|
| 5s output (default) | 5s | $0.90 | $1.00 | $1.15 |
| 10s output | 10s | $1.80 | $2.00 | $2.30 |
| 30s output | 30s | $5.40 | $6.00 | $6.90 |
| 5s output + 8s reference video | 13s | $2.34 | $2.60 | $2.99 |
| 5s output + 15s reference videos | 20s | $3.60 | $4.00 | $4.60 |

With `duration: "auto"` the run is pre-authorized at 30 seconds, then settled on the length that was actually generated. Failed runs are not charged. Output files are copied to SandBase storage and the returned URL stays valid for 7 days; download anything you need to keep.

#### Prompting

- Assign each reference a role: “Image 1 is the courier, Video 1 sets the camera move, Audio 1 is her voice.”
- Numbering is per type and follows array order, so reordering `images` changes what “Image 2” means.
- Put spoken lines in quotes and name the speaker.
- Wan 3.0 Pro has no negative prompt; say what you want rather than what to avoid.

#### Limits

- Reference video length plus output length must stay within 30 seconds.
- Up to 10 images, 5 videos (15s total), and 5 audio clips (15s total) per request.
- Older integrations that send `media: [{type, url}]` are still accepted and treated as reference images; new integrations should use the fields above.

#### Also on SandBase

Other Wan 3.0 Pro endpoints: `alibaba/wan/3.0/pro/text-to-video`, `alibaba/wan/3.0/pro/image-to-video`, `alibaba/wan/3.0/pro/reference-to-video`, `alibaba/wan/3.0/pro/video-edit`, `alibaba/wan/3.0/pro/video-extend`.