OpenAI Launches GPT-Image-1: Bringing ChatGPT‑Style Image Generation to Developers

OpenAI recently released the GPT‑Image‑1 model, a native multimodal large language model that differs from earlier dedicated image generators like DALL·E 2 and DALL·E 3 by supporting both image generation and editing in a single API.

Model Features

Users can customize output by specifying quality, image dimensions, file format, compression level, and whether a transparent background is needed. The model accepts a minimum image side length of 1024 px and supports three resolutions: 1024×1024, 1024×1536, and 1536×1024.

Accessing the API

The quickest way to try the capability is via the ChatGPT platform, but developers who need programmatic access should use the OpenAI developer portal. Some developers must complete organization verification, which is indicated by a “Verify organization” button in the settings.

After verification, the Playground’s Image tab can be used to generate sample images.

Prompt: An image of an astronaut riding a horse on the moon

The response shows the used input and output token counts and image metadata.

Generating Images (Generations)

Example JavaScript code for a generation request:

import OpenAI from "openai";
const openai = new OpenAI({ apiKey: process.env.OPENAI_API_KEY });
const response = await openai.images.generate({
  model: "gpt-image-1",
  prompt: "An image of an astronaut riding a horse on the moon",
  n: 1,
  size: "1024x1024",
  quality: "high"
});

The API returns a Base64‑encoded image that is valid for about 60 minutes.

Editing Images (Edits)

Editing requires one or more source images and a prompt. Example using four product images:

import fs from "fs";
import OpenAI, { toFile } from "openai";
const client = new OpenAI();
const imageFiles = ["bath-bomb.png", "body-lotion.png", "incense-kit.png", "soap.png"];
const images = await Promise.all(
  imageFiles.map(async (file) => await toFile(fs.createReadStream(file), null, { type: "image/png" }))
);
const rsp = await client.images.edit({
  model: "gpt-image-1",
  image: images,
  prompt: "Create a lovely gift basket with these four items in it"
});
const image_base64 = rsp.data[0].b64_json;
const image_bytes = Buffer.from(image_base64, "base64");
fs.writeFileSync("basket.png", image_bytes);

Mask‑based editing is also supported; the mask’s transparent area is replaced while the black area remains unchanged.

Pricing

Charges are token‑based and differ for text input, image input, and image output:

Text input: $5 per 1 M tokens

Image input: $10 per 1 M tokens

Image output: $40 per 1 M tokens

In practice this translates to roughly $0.02 for a low‑quality image, $0.07 for medium quality, and $0.19 for a high‑quality square image.

Impact and Early Adoption

Several platforms—Adobe, Canva, Figma, GoDaddy, Airtable, and HeyGen—have already begun experimenting with GPT‑Image‑1, with some mentioned in OpenAI’s official blog. The author expects more integrations in the coming weeks.

Conclusion

The release fulfills a long‑awaited demand for a ChatGPT‑style image generation API, offering fast, versatile generation and editing capabilities that are already being adopted by major creative tools.