OpenAI just released a major upgrade to ChatGPT's image↗ capabilities. The new version, powered by GPT Image 1.5, generates images up to four times faster and makes edits that actually preserve what you care about.

The update is rolling out today for all ChatGPT users. It's also available through the API for developers who want to build image generation into their own applications.

Edits that don't mess everything up

Here's what makes this different. When you ask ChatGPT to edit an uploaded photo, it now changes only what you specifically request. Everything else stays consistent.

Lighting remains the same. Composition stays intact. If there are people in the image, their appearance doesn't randomly shift between edits.

This sounds basic, but previous versions struggled with this. You'd ask for a simple change and get back an image where everything looked slightly different. That's frustrating when you just wanted one specific adjustment.

What you can actually do with it

The model handles different types of edits effectively. You can add elements, subtract them, combine multiple images, blend styles, or transpose objects.

Want to try different hairstyles or clothing on a photo? The model keeps your face and body consistent while changing only what you asked for. Need to add text or adjust layout? It handles that while preserving important visual details.

Creative transformations work for both simple concepts and intricate ideas. OpenAI included preset styles and options, so you don't always need to write detailed prompts. Just select what you want and the model applies it.

Speed matters ~

Generating images four times faster changes how you use the tool. You can experiment more, try different variations quickly, and iterate without waiting around.

This matters especially when you're exploring creative ideas. Faster generation means more experimentation, which usually leads to better final results.

The new Images model is available today for all ChatGPT users. The updated Images experience within ChatGPT is rolling out for most users now, with Business and Enterprise access coming later.

If you use the API, GPT Image↗ 1.5 is available immediately for integration into your applications.