OpenAI Tipped to Upgrade ChatGPT’s Image Generator With Preset Styles

ChatGPT users could soon get an upgraded image generator, which borrows a feature from the Sora app. As per multiple leaks, OpenAI appears to be testing a version two of its native image generator, and is now letting users generate images in preset styles. If true, this will save users the hassle of detailing out the image style along with their text prompts and ensure higher image consistency. The new feature has popped up for several users across the chatbot's web client and mobile apps, and could be rolled out globally soon.

ChatGPT Could Soon Let Users Choose Image Styles

In a post on X (formerly known as Twitter), Tibor Blaho, the lead engineer at AIPRM, shared a screenshot of a new banner on the ChatGPT web. Dubbed “ImageGenV2Banner,” the banner mentions, “Make images with ChatGPT - Feature your friends, faster image generation, and more consistent editing.”

The banner was spotted on the enterprise account of ChatGPT, and others have also confirmed spotting similar banners. Separately, a user in South Korea posted about seeing a pop-up message in the beta version of the ChatGPT app. The message, as per the screenshot, mentioned trending styles that users can now choose before generating an image.

Notably, Styles as a feature was first added to OpenAI's Sora app, which can generate videos from text and images. It appears that the AI giant is now bringing the capability to ChatGPT. Based on the screenshots, users will now see visual cards for the trending art styles. The list is likely prepared based on the styles users are requesting more of.

While the leaks did not talk about the functionality of the feature, it is expected that once users select a style, their prompts will be converted accordingly. As mentioned above, this will eliminate the need for users to describe the art style in painstaking detail. Some of the suggested styles shown in the screenshot include Crayon, Chibi, Nightcam, Blockart, and more.

However, it should be noted that the information shared here has been collated from screenshots shared by tipsters and users, but the company has not officially announced the feature. Until OpenAI formally releases it, it cannot be said for sure that the company is planning to introduce the feature.