GPT-4o Image Generation API

Use Cases

gpt-4o native image generation for text-to-image, image-to-image and conversational-image-editing!!

Far more advanced text rendering than Midjourney or Dalle-3

Pricing

$0.05 per successful gpt-4o-image generation request

$0.02 per generation IF chatgpt randomly downgrades the generation model to Dalle-3

$0.001 per request IF the completion response does not contains an image(in most case that means you violated openai content policy, or rejected by the model)

$0 if the response was 4xx or 5xx error

Important Notice

It uses LLM API

It uses a new model named gpt-4o-image

It can only be used in stream mode. How to handle stream response

It supports multiple images as input

The outputs are slightly different than the legacy gpt-4o-image-preview model, see below for more detailed examples.

Request Examples

Simple Text to Image

curl --location 'https://api.piapi.ai/v1/chat/completions' \
--header 'Content-Type: application/json' \
--header 'Authorization: Bearer {your-api-key}' \
--data '{
    "model": "gpt-4o-image",
    "messages": [
        {
            "role": "user",
            "content": [
                {
                    "type": "text",
                    "text": "draw a cyberpunk city"
                }
            ]
        }
    ],
    "stream": true
}'

Simple Image to Image

curl --location 'https://api.piapi.ai/v1/chat/completions' \
--header 'Content-Type: application/json' \
--header 'Authorization: Bearer {your-api-key}' \
--data '{
    "model": "gpt-4o-image",
    "messages": [
        {
            "role": "user",
            "content": [
                {
                    "type": "image_url",
                    "image_url": {
                        "url": "https://images.ctfassets.net/kftzwdyauwt9/21orfxKx8HXXGKH8cTOq60/1eb34535ddce9c9e91fab0fad77bc158/minnias_cat_input.png?w=640&q=90&fm=webp"
                    }
                },
                {
                    "type": "text",
                    "text": "Give this cat a detective hat and a monocle"
                }
            ]
        }
    ],
    "stream": true
}'

Multiple Images Input

curl --location 'https://api.piapi.ai/v1/chat/completions' \
--header 'Content-Type: application/json' \
--header 'Authorization: Bearer {your-api-key}' \
--data '{
    "model": "gpt-4o-image",
    "messages": [
        {
            "role": "user",
            "content": [
                {
                    "type": "image_url",
                    "image_url": {
                        "url": "https://img.moegirl.org.cn/common/thumb/5/53/Rei_Ayanami.jpg/280px-Rei_Ayanami.jpg"
                    }
                },
                {
                    "type": "image_url",
                    "image_url": {
                        "url": "https://moegirl.uk/images/e/ef/OP_C057_asuka.jpg"
                    }
                },
                {
                    "type": "text",
                    "text": "merge two image"
                }
            ]
        }
    ],
    "stream": true
}'

Multiple Rounds Conversation

Different from gpt-4o-image-preview, gpt-4o-image does not have multi rounds feature. It's a WIP feature for gpt-4o-image.

Legacy gpt-4o-image-preview Model

It will remain exactly the same as before, until 2025-05-07

After 2025-05-07, all requests for gpt-4o-image-preview will be treated as gpt-4o-image, including inputs, outputs, pricings.

GPT-4o Image Generation API

Use Cases#

Pricing#

Important Notice#

Request Examples#

Simple Text to Image#

Simple Image to Image#

Multiple Images Input#

Multiple Rounds Conversation#

Legacy gpt-4o-image-preview Model#