GPT Image 2.5 Precise

GPT Image 2.5 Precise is the editing variant of the GPT Image 2.5 pair, tuned to apply a change to images you supply rather than invent a frame from scratch. Attach up to four references with a prompt describing the change, and it holds the composition you gave it instead of reimagining it, which is the usual failure of editing through a generation model. It also generates from text alone if you send no image. Billing follows the tokens it renders plus the pixels you send in, so cost tracks the size and detail you ask for, and the listed price is one standard square image at the high detail level.

Constraints & limits

The gateway auto-clamps requests to the model's limits where possible.

Quick start

This sample hits the image generation endpoint. Change base_url and your key, then run it.

from openai import OpenAI

client = OpenAI(
    api_key="tl-xxxxxxxxxxxxxxxxxxxxxxxx",
    base_url="https://api.thalam.ai/v1",
)

response = client.images.generate(
    model="openai/gpt-image-2.5-precise",
    prompt="A futuristic city skyline at sunset.",
    size="1024x1024",
    n=1,
)

print(response.data[0].url)