# GPT Image 2.5 Sunburst - Core.Today AI API

> OpenAI's most capable image model, built for workflows where editing precision matters most. Generate and edit images with close instruction following — it keeps identity, composition and lighting while changing only what you ask. Same pricing as GPT Image 2.5 Flare, which is faster.

- **Provider**: OpenAI
- **Model ID**: openai/gpt-image-2.5-sunburst
- **Category**: Image Generation
- **Credits**: 580 per image (default quality=auto) × number_of_images — low 28 / medium 109 / high 298 / xhigh·auto 581 / max 1,163 credits per image
- **Speed**: Medium
- **Quality**: Ultra

## Features
- OpenAI's most capable image model — highest editing precision in the 2.5 line
- Targeted edits that preserve identity, composition and lighting
- Text-to-image and image editing in one model — just add input_images
- Dense text, small lettering and complex layouts rendered sharply
- Six quality tiers: low / medium / high / xhigh / max / auto
- Up to 10 images per request (billed per image)

## Use Cases
- Placing people into new scenes while preserving their likeness
- Interior edits — swap furniture without changing the camera angle
- Product photo retouching, object removal and lighting changes
- Translating text inside images and localizing marketing assets
- Multi-page illustrations with consistent characters

## 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 - What you want to generate, or how to edit the input images. Put exact text in "quotes".

### Optional
- **input_images**: array - Reference images for editing or composition. Pass one image to edit it, or several to combine subjects and styles. Processed at high fidelity automatically.
- **quality**: string (default: auto) - Quality tier. Higher tiers take longer and cost more per image: low 28, medium 109, high 298, xhigh 581, max 1,163 credits. auto is billed at the xhigh rate.
  Options: low, medium, high, xhigh, max, auto
- **aspect_ratio**: string (default: 1:1) - Named ratio, explicit size, or auto to let the model pick. Sizes above 2560x1440 are experimental and may give more variable results.
  Options: 1:1, 3:2, 2:3, 4:3, 3:4, 16:9, 9:16, auto, 1024x1024, 1536x1024, 1024x1536, 1536x1152, 1152x1536, 2048x2048, 2048x1152, 1152x2048, 3840x2160, 2160x3840
- **number_of_images**: integer (default: 1) - Number of images to generate (1-10). Billed per image.
- **background**: string (default: auto) - Background mode. For a transparent background use png or webp output.
  Options: auto, transparent, opaque
- **output_format**: string (default: webp) - Output image format
  Options: png, jpeg, webp
- **output_compression**: integer (default: 90) - Compression level (0-100) for webp/jpeg output
- **moderation**: string (default: auto) - Content moderation level — low applies less strict filtering
  Options: auto, low
- **user_id**: string - An optional unique identifier representing your end-user. This helps OpenAI monitor and detect abuse.
- **openai_api_key**: string - Your own OpenAI API key (optional — leave empty to use the gateway)

## Examples

### Photorealistic Close-up
Photo language (depth of field) at the high quality tier
```json
{
  "model": "openai/gpt-image-2.5-sunburst",
  "input": {
    "prompt": "A golden retriever puppy wearing tiny sunglasses on a sunny beach, photorealistic, shallow depth of field",
    "quality": "high",
    "aspect_ratio": "1:1"
  }
}
```

### Precise Photo Edit
Change one element and lock everything else
```json
{
  "model": "openai/gpt-image-2.5-sunburst",
  "input": {
    "prompt": "Replace the grey sofa with a light blue velvet sofa. Keep the camera angle, room layout, lighting and every other object exactly the same.",
    "input_images": [
      "https://example.com/living-room.jpg"
    ],
    "quality": "high",
    "aspect_ratio": "auto"
  }
}
```

### Composite Multiple References
Combine a person and a product into one scene
```json
{
  "model": "openai/gpt-image-2.5-sunburst",
  "input": {
    "prompt": "Place the person from the first image holding the handbag from the second image, walking down a Paris street in morning light. Preserve the person's face and the bag's exact design.",
    "input_images": [
      "https://example.com/model.jpg",
      "https://example.com/handbag.png"
    ],
    "quality": "high",
    "aspect_ratio": "2:3"
  }
}
```

### Translate Text in an Image
Localize a poster while keeping its design
```json
{
  "model": "openai/gpt-image-2.5-sunburst",
  "input": {
    "prompt": "Translate all English text on this poster into Korean. The headline should read \"여름 한정 세일\". Keep the layout, fonts' weight and colors unchanged.",
    "input_images": [
      "https://example.com/poster.png"
    ],
    "quality": "high",
    "output_format": "png"
  }
}
```

## 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
- Always set quality — the default auto is billed like xhigh (581 credits per image)
- When editing, say what must not change: "change only the lighting, preserve the face, pose and clothing"
- Be specific: instead of "make it better", say "add soft coastal daylight"
- Put exact on-image text in "quotes" and describe the typography
- With several input_images, refer to them by order ("the first image", "the second image")
- For fast, high-volume generation use openai/gpt-image-2.5-flare — same parameters and pricing