openai/gpt-image-2/image-to-image

Image to Image

OpenAI GPT Image 2 makes fine-grained edits to your images, from rewriting in-image text to restyling props and scenes. Combine up to 16 reference images, add a mask for targeted changes, pick low to high quality and get up to 4 results per call.

Try it Get API keyOpenAPI llms.txt

Prompt *

317 / 32,000

Image URLs *

1 *

Mask URL

Image size

More settings (4)

Background

Number of Images

Output Format

Quality

ResultExampleSign in to try

Billed per token · about $0.15 per image at default settings · $0.151 reserved per request

Example — generated from this input

Turn this into a winter wedding invitation: change the date line to read "DECEMBER 12, 2027 · ZERMATT", replace the eucalyptus with snowy pine sprigs and small pinecones, and make the ribbon deep burgundy velvet. Keep "Emma & Lucas", the gold script, the card design, the marble and the camera angle exactly the same.

Pricing

Text input /M: $5.00
Image input /M: $8.00
Image output /M: $30.00
Cache read /M: $2.00

About $0.15 per image at default settings (high quality, 1 image); higher quality, larger sizes and more images use more tokens.

Each request temporarily reserves $0.151; you are charged for actual usage, usually within a minute: any unused amount is released and any usage above it is charged to your balance, even if that takes it below zero; while your balance is below zero, new requests are paused until a top-up covers it.

Failed generations are not charged.

API

Submit generation requests with this model ID.

POST/v1/model-tasks

openai/gpt-image-2/image-to-image

Request body

{
  "model": "openai/gpt-image-2/image-to-image",
  "input": {
    "prompt": "Turn this into a winter wedding invitation: change the date line to read \"DECEMBER 12, 2027 · ZERMATT\", replace the eucalyptus with snowy pine sprigs and small pinecones, and make the ribbon deep burgundy velvet. Keep \"Emma & Lucas\", the gold script, the card design, the marble and the camera angle exactly the same.",
    "background": "auto",
    "image_size": "auto",
    "image_urls": [
      "https://pub-a7287c8d5de944d984e22a5bd3ab9bbe.r2.dev/uploads/capability-demo-4ffa20a3b8f2c3e00ceb12f6d50e9559.png"
    ],
    "num_images": 1,
    "output_format": "png",
    "quality": "high"
  }
}

Input schema

Parameter	Type	Requirement	Default
image_urls	array	Required	—
prompt	string	Required	—
background	string	Optional	auto
image_size	object \| string	Optional	auto
mask_url	string	Optional	—
num_images	integer	Optional	1
output_format	string	Optional	png
quality	string	Optional	high

Response

The API returns a result or a pollable asynchronous task. See the examples and Schema below for the response mode and result fields.

Asynchronous task

Submit response

HTTP 202

/v1/model-tasksapplication/json

{
  "id": "task_example",
  "model": "openai/gpt-image-2/image-to-image",
  "status": "queued"
}

Status response

HTTP 200

/v1/model-tasks/{id}application/json

{
  "id": "task_example",
  "model": "openai/gpt-image-2/image-to-image",
  "status": "completed",
  "output": {
    "images": [
      {
        "url": "https://api.musespan.com/v1/model-tasks/task_example/artifacts/0"
      }
    ]
  }
}

File download

HTTP 200

/v1/model-tasks/{id}/artifacts/{index}Generated image (e.g. image/png)

When the task completes, file URLs in output point to this endpoint, numbered from 0. Send the same API key (Authorization header) when downloading.

Error response

HTTP 400

/v1/model-tasksapplication/json

{
  "error": {
    "code": "invalid_input",
    "message": "The input does not match this model's input schema. Check the model's API documentation.",
    "type": "invalid_request_error"
  }
}

Code examples

Async submission returns a task ID. Poll until status is completed, then retrieve the generated content from the result. Get your API key in the console.

# Use a new directory for each new generation. Rerunning here continues this task.
set -eu
test -s model-task-id.txt || python3 -c "import uuid; print(uuid.uuid4())" > model-task-id.txt
REQUEST_ID=$(cat model-task-id.txt)
# Idempotent: resending the same bytes with the same ID never charges twice.
curl --fail-with-body --retry 3 --retry-connrefused --request POST 'https://api.musespan.com/v1/model-tasks' \
  --header "Authorization: Bearer $MUSESPAN_API_KEY" \
  --header "X-Request-Id: $REQUEST_ID" \
  --header "Content-Type: application/json" \
  --data '{"model":"openai/gpt-image-2/image-to-image","input":{"prompt":"Turn this into a winter wedding invitation: change the date line to read \"DECEMBER 12, 2027 · ZERMATT\", replace the eucalyptus with snowy pine sprigs and small pinecones, and make the ribbon deep burgundy velvet. Keep \"Emma & Lucas\", the gold script, the card design, the marble and the camera angle exactly the same.","background":"auto","image_size":"auto","image_urls":["https://pub-a7287c8d5de944d984e22a5bd3ab9bbe.r2.dev/uploads/capability-demo-4ffa20a3b8f2c3e00ceb12f6d50e9559.png"],"num_images":1,"output_format":"png","quality":"high"}}'
# Poll every 5 seconds for up to 25 minutes until status is completed or failed.
# Still queued (with queue_position) 5 minutes after submission: failed with queue_timeout, not charged.
# No result 15 minutes after acceptance (or no confirmation 15 minutes after submission): failed with generation_timeout, not charged.
# Any failure is final: generate again in a new directory (a new X-Request-Id).
# A 404 means the POST did not arrive: rerun this script to resend it with the same ID.
curl --fail-with-body 'https://api.musespan.com/v1/model-tasks/'"$REQUEST_ID"'' --header "Authorization: Bearer $MUSESPAN_API_KEY"