Following the recent release of GPT-5.2, today OpenAI is unveiling the new and revamped ChatGPT Images, which is four times faster at generating images. This has obviously been developed to compete head-on with Google’s Nano Banana Pro, in the same way in which GPT-5.2 is a direct response to Gemini 3 Pro.
ChatGPT Images “adheres to your intent more reliably”, OpenAI’s press release says, when you ask it for edits, “down to the small details”. It changes only what you ask while keeping everything else intact across inputs, outputs, and subsequent edits.
This should unlock “results that match your intent”, allowing ChatGPT to “act as a creative studio in your pocket, capable of both practical edits and expressive reimaginings”. The new model is better at adding, subtracting, combining, blending, and transposing, “so you get the changes you want without losing what makes the image special”.
Within the new ChatGPT Images feature you’ll find preset styles and ideas you can pick without needing to enter a written prompt. The model also follows instructions more reliably, enabling more precise edits and more intricate original compositions, “where relationships between elements are preserved as intended”. Text rendering should be improved too, as well as rendering many small faces.
The new ChatGPT Images is rolling out now to all users globally.
