Images in ChatGPT are getting a major update

OpenAI on Tuesday unveiled a new image engine for ChatGPT that it says takes a key step forward in rendering text and handling more complex requests.

Why it matters: When a capable new model pops, it can lead to a viral moment and a surge in app downloads.

Driving the news: OpenAI said ChatGPT Images 2.0 supports a wide range of aspect ratios and comes in both standard and a "thinking" mode with built-in reasoning.

• All users will have access to the standard version of the image model, while the thinking mode is reserved for paid subscribers.
• The new models will also be available to developers via an API.

Yes, but: All that extra thinking can mean images take longer for ChatGPT to produce.

The big picture: The lead for any one model is often a short one, as new models catch up and often overtake the leader.

• Google scored significant attention last year when it https://www.axios.com/2025/08/26/nano-banana-google-ai-images" target="_blank">launched Nano Banana.
• OpenAI had a https://www.axios.com/2025/03/31/chatgpt-image-altman-openai-users" target="_blank">big bump of its own a few months earlier when it debuted a more capable model, with Studio Ghibli-style images going particularly viral.

What they're saying: OpenAI said it can't predict what will take off this time, but expects the new model to produce its own round of viral images.

• "We believe that we are going to have another moment here," OpenAI product manager Adele Li said during a briefing with reporters.

More importantly, though, OpenAI says the new model is well suited to a wide range of professional work including ads, posters and mock-ups.

• "It's not just a tool for making beautiful pictures," Li said. "It's a creative assistant."