OpenAI is releasing an update to its artificial intelligence image-generating software that it says will let users create accurate, complex charts and scientific diagrams, part of a bid by the company to make its technology more appealing to professionals.
ChatGPT Images 2.0 is set to roll out Tuesday through OpenAI’s flagship chatbot and Codex AI coding assistant. The company said the product will be better at following a user’s instructions when generating an image and will be able to include more details. It can also spit out visuals that more faithfully reflect a slew of styles and render text in multiple languages, the company said.
The ChatGPT maker has moved to streamline its product offerings in recent weeks to compete for business customers against rival Anthropic PBC and pave the way for a possible initial public offering as soon as this year. As part of that effort, the company shuttered its Sora AI video generator. But OpenAI sees continued value in offering an AI image tool.
In a briefing with reporters this week, Adele Li, who leads the product team for ChatGPT Images, said the functionality is “core” to what users want to do with AI. Hundreds of millions of weekly active ChatGPT users make images with the company’s software, the company said, and more than 1 billion images are created each week with the chatbot.
Li said OpenAI has made progress specifically on generating images that have structured layouts and complicated scientific diagrams in particular. These advances could lead to more practical uses of the technology by educators, scientists and parents, among others, she said.
OpenAI’s new image model can also spend more time computing before responding to a user, which the company said allows the software to find details online and ruminate more over how to construct an image. This feature will be available to paid users, the company said.
Metz writes for Bloomberg.