Where can I generate images with GPT Image 2.6?

In ChatGPT, in Codex, through the API, or in the box on this page.

ChatGPT and Codex have had the model worldwide since 4 November 2026. Developers call it through the images API, with GPT-Image-2.6 Flare or GPT-Image-2.6 Sunburst. Image26 is a separate page for the same models. It returns a file when an API key is set on the server. It is not an OpenAI product, and it is not affiliated with OpenAI.

Pick the job first

A still image covers a poster, a product shot, or a scene. Flare is the quick way to see if the idea works. A GIF covers a reaction, a product loop, or a storyboard you can watch before anyone books a video shoot. An SVG covers a logo or icon made of flat shapes. A rough sketch can become a character you reuse.

On this page, choose Image, GIF, or SVG, then write the subject, the setting, and anything that must stay put. The sizes on offer are square 1024×1024, wide 1536×1024, portrait 1024×1536, and poster 1024×1280. Each side is a multiple of 16.

What this page sends

The box sends the prompt, the model, the shape, the detail, and the planning level. It does not stream partial images. A GIF request from here is 9 frames at 200 milliseconds. An SVG request is a single file. If you want the call inside your own site, use the API and keep the key on your server.