> ## Documentation Index
> Fetch the complete documentation index at: https://docs.arouter.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Image Generation

> Generate images from text prompts using models like DALL-E 3, Stable Diffusion, and Flux. ARouter routes image generation requests through a single unified endpoint.

ARouter supports image generation via the OpenAI-compatible `/v1/images/generations` endpoint. Use any supported image generation model with a single API key.

## Quick Start

```bash theme={null}
curl https://api.arouter.ai/v1/images/generations \
  -H "Authorization: Bearer lr_live_xxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/dall-e-3",
    "prompt": "A futuristic city skyline at sunset, photorealistic",
    "n": 1,
    "size": "1024x1024"
  }'
```

<Tabs>
  <Tab title="Python">
    ```python theme={null}
    from openai import OpenAI

    client = OpenAI(
        base_url="https://api.arouter.ai/v1",
        api_key="lr_live_xxxx",
    )

    response = client.images.generate(
        model="openai/dall-e-3",
        prompt="A futuristic city skyline at sunset, photorealistic",
        n=1,
        size="1024x1024",
        quality="hd",
    )

    print(response.data[0].url)
    ```
  </Tab>

  <Tab title="Node.js">
    ```typescript theme={null}
    import OpenAI from "openai";

    const client = new OpenAI({
      baseURL: "https://api.arouter.ai/v1",
      apiKey: "lr_live_xxxx",
    });

    const response = await client.images.generate({
      model: "openai/dall-e-3",
      prompt: "A futuristic city skyline at sunset, photorealistic",
      n: 1,
      size: "1024x1024",
      quality: "hd",
    });

    console.log(response.data[0].url);
    ```
  </Tab>

  <Tab title="cURL">
    ```bash theme={null}
    curl https://api.arouter.ai/v1/images/generations \
      -H "Authorization: Bearer lr_live_xxxx" \
      -H "Content-Type: application/json" \
      -d '{
        "model": "openai/dall-e-3",
        "prompt": "A futuristic city skyline at sunset, photorealistic",
        "n": 1,
        "size": "1024x1024",
        "quality": "hd"
      }'
    ```
  </Tab>
</Tabs>

***

## Request Parameters

| Parameter         | Type      | Description                                                                                   |
| ----------------- | --------- | --------------------------------------------------------------------------------------------- |
| `model`           | `string`  | Image generation model ID (see [Supported Models](#supported-models))                         |
| `prompt`          | `string`  | Text description of the desired image. Max 4,000 characters for DALL-E 3, 1,000 for DALL-E 2. |
| `n`               | `integer` | Number of images to generate (default `1`). DALL-E 3 only supports `n=1`.                     |
| `size`            | `string`  | Image dimensions (model-specific, see below)                                                  |
| `quality`         | `string`  | `"standard"` or `"hd"` (DALL-E 3 only, default `"standard"`)                                  |
| `style`           | `string`  | `"vivid"` (hyper-real) or `"natural"` (more realistic, DALL-E 3 only)                         |
| `response_format` | `string`  | `"url"` (default) or `"b64_json"`                                                             |

***

## Response Format

```json theme={null}
{
  "created": 1234567890,
  "data": [
    {
      "url": "https://...",
      "revised_prompt": "A photorealistic futuristic city skyline at sunset..."
    }
  ]
}
```

| Field            | Description                                                                                  |
| ---------------- | -------------------------------------------------------------------------------------------- |
| `url`            | Temporary URL to the generated image (expires after 1 hour)                                  |
| `b64_json`       | Base64-encoded image data (when `response_format: "b64_json"`)                               |
| `revised_prompt` | DALL-E 3 may revise your prompt for better results. This field shows the actual prompt used. |

***

## Image Sizes

### DALL-E 3

| Size        | Aspect Ratio     |
| ----------- | ---------------- |
| `1024x1024` | Square (1:1)     |
| `1792x1024` | Landscape (16:9) |
| `1024x1792` | Portrait (9:16)  |

### DALL-E 2

| Size        |
| ----------- |
| `256x256`   |
| `512x512`   |
| `1024x1024` |

***

## Supported Models

Use `GET /v1/models?output_modalities=image` to query the current list of image generation models.

| Model                                | Provider     | Resolution      | Style Control |
| ------------------------------------ | ------------ | --------------- | ------------- |
| `openai/dall-e-3`                    | OpenAI       | Up to 1792×1024 | `style` param |
| `openai/dall-e-2`                    | OpenAI       | Up to 1024×1024 | —             |
| `stability/stable-diffusion-3-large` | Stability AI | Up to 1024×1024 | —             |
| `black-forest-labs/flux-1-pro`       | Replicate    | Flexible        | —             |
| `black-forest-labs/flux-1-schnell`   | Replicate    | Flexible        | —             |

<Note>
  Model availability may vary. Always query `GET /v1/models?output_modalities=image` for the current list.
</Note>

***

## Advanced Examples

### High-Quality Portrait

```python theme={null}
response = client.images.generate(
    model="openai/dall-e-3",
    prompt="Professional headshot of a software engineer, studio lighting, sharp focus",
    size="1024x1792",
    quality="hd",
    style="natural",
    response_format="b64_json",
)

import base64
image_bytes = base64.b64decode(response.data[0].b64_json)
with open("portrait.png", "wb") as f:
    f.write(image_bytes)
```

### Batch Generation with DALL-E 2

```python theme={null}
response = client.images.generate(
    model="openai/dall-e-2",
    prompt="A friendly robot assistant helping with paperwork",
    n=4,
    size="512x512",
)

for i, image in enumerate(response.data):
    print(f"Image {i+1}: {image.url}")
```

***

## Routing to Specific Providers

Use the `provider` object to control which provider handles your image generation request:

```json theme={null}
{
  "model": "openai/dall-e-3",
  "prompt": "A mountain landscape at dawn",
  "provider": {
    "order": ["OpenAI"]
  }
}
```

See [Provider Routing](/en/provider-routing) for full routing options.

***

## Pricing

Image generation is billed per image, not per token. Costs vary by model and resolution:

| Model                      | Size                   | Cost    |
| -------------------------- | ---------------------- | ------- |
| `openai/dall-e-3` standard | 1024×1024              | \$0.040 |
| `openai/dall-e-3` standard | 1024×1792 or 1792×1024 | \$0.080 |
| `openai/dall-e-3` HD       | 1024×1024              | \$0.080 |
| `openai/dall-e-3` HD       | 1024×1792 or 1792×1024 | \$0.120 |
| `openai/dall-e-2`          | 1024×1024              | \$0.020 |
| `openai/dall-e-2`          | 512×512                | \$0.018 |
| `openai/dall-e-2`          | 256×256                | \$0.016 |

Costs for other models (Stability AI, Flux) are charged based on the upstream provider rate. Check `usage.cost` in the response for the actual charge.

<Note>
  Image URLs returned by ARouter expire after 1 hour. Download and store images you intend to keep.
</Note>