Visual Leap: OpenAI’s latest image tool offers multiple options with a single prompt

 When the big AI firms launched their image-generating tools last year, social media was soon awash with viral trends featuring images cre­ated with these tools. There was the Studio Ghibli

trend, where people posted images of themselves created in the style of the famous animation studio by using OpenAI’s GPT-4o, and then there was the Nano Banana trend where the images uploaded were hyper-realistic figurines of themselves made using Google’s image generator.

OpenAI’s newly launched AI-powered image generator, called ChatGPT Images 2, hasn’t led to any such viral trend—at least not yet. But it repre­sents a new way of generating such visuals, going from instant interpretation to what OpenAI calls deliberate construction.

ChatGPT Images 2 has what the firm calls “rea­soning capabilities”, allowing the tool to search the

web to create multiple images from a single prompt.

In essence, the bot can now perform additional steps, such as searching the internet for recent information and references to create more sophisticated images than those gen­erated by earlier models. “Images 2.0 is a huge step forward—like going from GPT-3 to GPT-5 in one leap. The ability to create incredible new images, express creativity, and produce beautiful,

complex visuals is quite remarkable,” said Sam Altman, OpenAI’s CEO, at the launch of the update. Since the model can generate multiple images from a single prompt, it is ideally suited to create entire comic books or series of graphics. Many more options will probably open up in due course. The tool is also said to be good at generat­ing images with texts, something that was diffi­cult with previous iterations, including those in many foreign languages.