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

# GPT-Image-2

> Generate stunning AI images with OpenAI's GPT-Image-2 — a purpose-built image generation model with 13 aspect ratios, up to 16 reference images, and exceptional prompt understanding.

## Model Description

OpenAI's dedicated image generation model (`gpt-image-2`) delivers superior visual quality with broad aspect ratio support and multi-reference image fusion.

**Maximum Resolution:** 4K
**Pricing:** Starts at \$0.016 per image

### Key Capabilities

* **Purpose-Built for Images:** Designed exclusively for image generation, delivering superior visual quality compared to multimodal models.
* **13 Aspect Ratios:** From ultra-wide 21:9 to ultra-tall 9:21 — perfect for any platform or format.
* **16 Reference Images:** Upload up to 16 reference images for style guidance, character consistency, and multi-image fusion.
* **Accurate Text Rendering:** Industry-leading accuracy for text, logos, and typography within generated images.

### Use Cases

* **Product Photography:** Professional product shots and packaging designs with text overlays.
* **Marketing Creatives:** Ad banners, social media graphics, and promotional visuals at scale.
* **Brand Identity:** Consistent brand visuals using reference images for style continuity.

***

## Quick Start

### 1. Generating an Image with GPT-Image-2

```bash theme={null}
curl --request POST \
  --url https://api.veogen.studio/api/v1/generations \
  --header 'Authorization: Bearer <API_KEY>' \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "gpt-image-2",
    "prompt": "A luxury perfume bottle on black marble with dramatic studio lighting, the text \"AURORA\" engraved on the glass",
    "aspect_ratio": "3:2",
    "resolution": "4K"
  }'
```

<ResponseExample>
  ```json theme={null}
  {
    "data": {
      "id": "img_gptimg2_12345",
      "model": "gpt-image-2",
      "status": "pending",
      "created_at": "2026-04-22T12:00:00Z"
    }
  }
  ```
</ResponseExample>

### 2. Image-to-Image with Reference Images

```bash theme={null}
curl --request POST \
  --url https://api.veogen.studio/api/v1/generations \
  --header 'Authorization: Bearer <API_KEY>' \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "gpt-image-2",
    "prompt": "Transform this photo into a watercolor painting with soft pastel tones",
    "aspect_ratio": "1:1",
    "image_urls": [
      "https://example.com/reference-photo.jpg"
    ]
  }'
```

## Supported Aspect Ratios

| Ratio           | Description                      |
| --------------- | -------------------------------- |
| `1:1`           | Square (default)                 |
| `16:9` / `9:16` | Widescreen landscape / portrait  |
| `4:3` / `3:4`   | Standard landscape / portrait    |
| `3:2` / `2:3`   | Classic landscape / portrait     |
| `5:4` / `4:5`   | Near-square landscape / portrait |
| `2:1` / `1:2`   | Wide landscape / tall portrait   |
| `21:9` / `9:21` | Ultra-wide / ultra-tall          |

<Warning>
  Only ratio notation is supported (e.g. `16:9`). Passing pixel sizes like `1024x1024` will return an error.
</Warning>

## FAQ

<AccordionGroup>
  <Accordion title="How is GPT-Image-2 different from GPT-4o Image?">
    GPT-Image-2 is a dedicated image generation model with 13 aspect ratios (vs 3), support for 16 reference images (vs 1), and generally better visual quality. GPT-4o Image is a multimodal model that also generates images.
  </Accordion>

  <Accordion title="Do I need an OpenAI account to use this?">
    No. Access GPT-Image-2 through Veogen's API or playground without any OpenAI subscription. Just top up your USD balance and generate.
  </Accordion>

  <Accordion title="How many reference images can I use?">
    Up to 16 reference images per request. You can mix image URLs and base64-encoded images for style guidance and multi-image fusion.
  </Accordion>
</AccordionGroup>