# GPT Image 1 - Core.Today AI API

> OpenAI's natively multimodal image generation model, served through the LLM gateway via the OpenAI Images API. Accurate in-image text rendering, strong instruction following, and fine-grained style control, billed by tokens.

- **Provider**: OpenAI
- **Model ID**: gpt-image-1
- **Category**: Image Generation
- **Credits**: 40 per image (approx, 1024×1024)
- **Speed**: Medium
- **Quality**: Ultra

## Features
- Natively multimodal image generation (text and image inputs, image outputs)
- Accurate in-image text rendering
- Strong instruction following and style control
- Token-based billing: text input tokens (9,290 credits/1M) and image output tokens (74,320 credits/1M) are billed separately
- Served through the LLM gateway via the OpenAI Images API

## Use Cases
- Marketing graphics with legible embedded text
- Illustrations and concept art with precise style control
- Product visuals from detailed written specs
- UI mockups and diagrams described in natural language

## 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
- **model**: string - Model identifier
- **prompt**: string - A text description of the desired image

### Optional
- **size**: string (default: auto) - Output image size
  Options: 1024x1024, 1536x1024, 1024x1536, auto
- **quality**: string (default: auto) - Rendering quality. Higher quality produces more image output tokens and therefore costs more
  Options: low, medium, high, auto
- **background**: string (default: auto) - Set whether the background is transparent, opaque, or chosen automatically
  Options: auto, transparent, opaque
- **output_format**: string (default: png) - Output image format
  Options: png, jpeg, webp
- **n**: integer (default: 1) - Number of images to generate. Each image bills its own output tokens

## Examples

### Image Generation
Generate an image via the OpenAI Images API
```json
{
  "model": "gpt-image-1",
  "input": {
    "model": "gpt-image-1",
    "prompt": "A storefront window poster that reads \"GRAND OPENING - 50% OFF\", warm evening light, photorealistic",
    "size": "1024x1024",
    "quality": "high"
  }
}
```

## 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
- Works with the OpenAI SDK - set base_url to https://ai.api.core.today/llm/openai/v1 and use client.images.generate()
- Billing is token-based: text input tokens at 9,290 credits/1M plus image output tokens at 74,320 credits/1M - larger sizes and higher quality produce more output tokens
- The ~40 credits per image figure assumes a 1024×1024 output; treat it as an estimate, not a flat rate
- Use GPT Image 1 Mini for drafts and iteration, then rerun the final prompt here

## Documentation
https://platform.openai.com/docs/models/gpt-image-1