GPT Image

gpt-image-2 API

gpt-image-2 supports text-to-image generation and image editing. It is available through both the AIHubMix unified generation endpoint and OpenAI-compatible endpoints.

Scenario	Method	Path	Content-Type
Unified generation	POST	`/ai/v1/images/generations`	`application/json`
OpenAI-compatible generation	POST	`/v1/images/generations`	`application/json`
OpenAI-compatible editing	POST	`/v1/images/edits`	`multipart/form-data`

For new unified integrations, see Image Generation. The examples on this page use OpenAI-compatible endpoints for existing OpenAI SDK integrations.

Generation, editing, Base64 decoding, and the invalid-parameter flow on this page were verified with gpt-image-2. Before integration, query the Model Schema API, select an endpoint by path, and read its request.schema.

Prerequisites

Install or upgrade the OpenAI Python SDK and provide the API key through an environment variable:

pip install -U openai
export AIHUBMIX_API_KEY="YOUR_API_KEY"

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["AIHUBMIX_API_KEY"],
    base_url="https://aihubmix.com/v1",
)

Generate an image

This example generates one low-quality 1024x1024 PNG and saves the returned Base64 data.

import base64
import os
from pathlib import Path
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["AIHUBMIX_API_KEY"],
    base_url="https://aihubmix.com/v1",
)

result = client.images.generate(
    model="gpt-image-2",
    prompt="A simple green triangle centered on a plain white background",
    n=1,
    size="1024x1024",
    quality="low",
    output_format="png",
)

image_bytes = base64.b64decode(result.data[0].b64_json)
Path("gpt-image-2.png").write_bytes(image_bytes)
print(result.usage)

Equivalent HTTP request:

curl https://aihubmix.com/v1/images/generations \
  -H "Authorization: Bearer $AIHUBMIX_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "A simple green triangle centered on a plain white background",
    "n": 1,
    "size": "1024x1024",
    "quality": "low",
    "output_format": "png"
  }'

Edit an image

The editing endpoint accepts multipart/form-data. Put the source image in the current directory as input.png.

import base64
import os
from pathlib import Path
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["AIHUBMIX_API_KEY"],
    base_url="https://aihubmix.com/v1",
)

with Path("input.png").open("rb") as source_image:
    result = client.images.edit(
        model="gpt-image-2",
        image=source_image,
        prompt="Change the red circle to a blue square, keeping the white background",
        n=1,
        size="1024x1024",
        quality="low",
        output_format="png",
    )

image_bytes = base64.b64decode(result.data[0].b64_json)
Path("gpt-image-2-edited.png").write_bytes(image_bytes)
print(result.usage)

Equivalent HTTP request:

curl https://aihubmix.com/v1/images/edits \
  -H "Authorization: Bearer $AIHUBMIX_API_KEY" \
  -F "model=gpt-image-2" \
  -F "image=@input.png;type=image/png" \
  -F "prompt=Change the red circle to a blue square, keeping the white background" \
  -F "n=1" \
  -F "size=1024x1024" \
  -F "quality=low" \
  -F "output_format=png"

Confirmed parameters

Field	Generation	Editing	Description
`model`	Required	Required	`gpt-image-2`
`prompt`	Required	Required	Non-empty image description or editing instruction
`image`	-	Required	Source image to edit
`mask`	-	Optional	Mask that selects the editing region
`n`	Optional	Optional	From `1` to `10`; default `1`
`size`	Optional	Optional	`{width}x{height}` format
`quality`	Optional	Optional	`low`, `medium`, or `high`
`output_format`	Optional	Optional	`png` or `jpeg`

Do not copy parameters from other GPT Image versions into gpt-image-2. Use the request.schema for the selected endpoint as the source of truth.

Response example

Both compatible endpoints return an OpenAI image response. data[].b64_json contains Base64 image data, and usage reports token usage.

{
  "created": 1787719630,
  "background": "opaque",
  "data": [
    {
      "b64_json": "<base64-encoded-image>"
    }
  ],
  "output_format": "png",
  "quality": "low",
  "size": "1024x1024",
  "usage": {
    "input_tokens": 16,
    "input_tokens_details": {
      "image_tokens": 0,
      "text_tokens": 16
    },
    "output_tokens": 196,
    "output_tokens_details": {
      "image_tokens": 196,
      "text_tokens": 0
    },
    "total_tokens": 212
  }
}

Error response

An unsupported parameter returns a structured 4xx error. For example, quality: "ultra" returns HTTP 400:

{
  "error": {
    "message": "Invalid value: 'ultra'. Supported values are: 'low', 'medium', 'high', and 'auto'.",
    "type": "invalid_request_error",
    "param": "quality",
    "code": "invalid_value"
  }
}

Check error.param and error.code, then correct the request against the selected endpoint’s request.schema.

Getting Started

Gateway Features

Model Capabilities

Protocol Reference

Platform Management

gpt-image-2 API

Prerequisites

Generate an image

Edit an image

Confirmed parameters

Response example

Error response

​gpt-image-2 API

​Prerequisites

​Generate an image

​Edit an image

​Confirmed parameters

​Response example

​Error response

gpt-image-2 API

Prerequisites

Generate an image

Edit an image

Confirmed parameters

Response example

Error response