OpenAI seems to be on the verge of releasing a new image generation model, which will likely be called GPT-Image-2.
OpenAI appears to be teasing a new image model. “This is not a screenshot,” the company posted on X. It added an image of what looked like a screenshot of a ChatGPT chat. But the image likely isn’t a screenshot, but an image created by the new image model with a prompt asking it to create an image of a ChatGPT chat.
Several X users have shared similar “screenshots” that appear to be generated by OpenAI’s new model. A user shared a fake screenshot of OpenAI’s website announcing GPT 5.5.
Another shared an image of the New York Times in which OpenAI announcing GPT-Image-2 was the main headline.
In the past, AI companies have managed to get lots of new users with image models. OpenAI had started things off last year, when the Ghibli trend using its image generator had taken off and gone viral. Google had then responded with Nano Banana, which had similarly gone viral and got it plenty of new users.
And OpenAI could do with some traction at the moment. Its market share has fallen from nearly 80% to just over 50%, with Google Gemini and Claude both garnering significant traction. Its overall user-growth has declined too — as per Apptopia data, its app DAUs have been falling every month since October last year. But if this image model does have interesting new capabilities — maybe beyond the screenshots that OpenAI is sharing — it could give the company the fillip it needs as it looks to stave off the challenge from other frontier labs and hold on to its lead in the AI race.