GPT Image 2.5 generation

Model positioning

• Model ID: gpt-image-2.5, gpt-image-2.5-flare, gpt-image-2.5-sunburst.
• Capabilities: Text-to-image generation and image editing, with precise text rendering for posters, signs, interface assets and other images containing text.
• Response format: The response may include url or b64_json. Download or decode and save it immediately, according to the actual model response.

Versions and prices

• gpt-image-2.5: 1K everyday tier, $0.013/image, for everyday visuals, creative sketches and cost-sensitive tasks.
• gpt-image-2.5-flare: 4K fast tier, $0.02/image, prioritizing generation speed for quick previews and iteration.
• gpt-image-2.5-sunburst: 4K quality tier, $0.025/image, prioritizing visual quality and detailed editing for final deliverables.

Sunburst is the default for everyday visuals, drafts, high-resolution deliverables and image editing. Choose Flare when speed is explicitly prioritized or a high-resolution request needs lower cost; choose Image 2.5 for the lowest-cost requests without a high-resolution requirement. Explicit model, image count and budget choices take priority. The default model does not imply a default 4K canvas. See the Model Catalog for current prices and availability.

Image generation

POST https://api.yomiapi.com/v1/images/generations

Request parameters

Python example

First set the local YOMI_API_KEY environment variable. The script reads it directly without echoing the key.

import base64
import os
from urllib.request import urlopen
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["YOMI_API_KEY"],
    base_url="https://api.yomiapi.com/v1"
)

response = client.images.generate(
    model="gpt-image-2.5-sunburst",
    prompt="A minimalist coffee brand poster with the main heading MORNING BREW, warm colors",
    size="1024x1536"
)

item = response.data[0]
if item.b64_json:
    image_bytes = base64.b64decode(item.b64_json)
elif item.url:
    # Download without sending the Yomi API key or echoing the temporary URL.
    with urlopen(item.url, timeout=120) as image_response:
        image_bytes = image_response.read()
else:
    raise RuntimeError("No image data in the response")

with open("poster.png", "wb") as f:
    f.write(image_bytes)

cURL example

curl https://api.yomiapi.com/v1/images/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2.5-sunburst",
    "prompt": "A minimalist coffee brand poster with the main heading MORNING BREW, warm colors",
    "size": "1024x1536"
  }'

Image editing

The editing endpoint uploads the source image as multipart/form-data and modifies it according to the prompt. Common editing instructions:

• Add an element: "Add a cup of coffee to the table"
• Replace an element: "Replace the red car with a blue car"
• Remove an element: "Remove the people in the background and fill the area with natural scenery"

POST https://api.yomiapi.com/v1/images/edits

Python example

First set the local YOMI_API_KEY environment variable. The script reads it directly without echoing the key.

import base64
import os
from urllib.request import urlopen
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["YOMI_API_KEY"],
    base_url="https://api.yomiapi.com/v1"
)

with open("input.png", "rb") as source_image:
    response = client.images.edit(
        model="gpt-image-2.5-sunburst",
        image=source_image,
        prompt="Change the scene to nighttime, add a neon effect to the shop sign, and preserve the subject and text",
        size="1024x1536",
        n=1
    )

item = response.data[0]
if item.b64_json:
    image_bytes = base64.b64decode(item.b64_json)
elif item.url:
    # Download without sending the Yomi API key or echoing the temporary URL.
    with urlopen(item.url, timeout=120) as image_response:
        image_bytes = image_response.read()
else:
    raise RuntimeError("No image data in the response")

with open("edited.png", "wb") as f:
    f.write(image_bytes)

Billing and recommendations

• Billing: Image-model billing methods and unit prices depend on the Model Catalog and actual charges under Usage.
• Text in images: Put the exact text in quotation marks in your prompt, and specify font style and placement.
• Batch generation: Validate the prompt with one image before generating a batch. Retry 429 rate-limit errors with exponential backoff.