OpenAI continues on its ‘code red’ warpath with new image generation model

OpenAI is rolling out a brand new model of ChatGPT Photos that guarantees higher instruction-following, extra exact enhancing, and as much as 4x sooner picture era speeds.

The brand new mannequin, dubbed GPT Picture 1.5, is obtainable beginning Tuesday to all ChatGPT customers and by way of the API. It’s the newest escalation within the competitors with Google’s Gemini after OpenAI CEO Sam Altman final month declared a “code purple” in a leaked inside memo. The memo detailed OpenAI’s plans to regain its place because the AI chief after Google had begun to take market share following the discharge of Gemini 3, its newest flagship mannequin, and Nano Banana Professional, the most recent model of Google’s viral picture generator — each of which have topped the LMArena leaderboard throughout a number of benchmarks. 

Google maintains its lead even after OpenAI responded to its success final week with the launch of GPT-5.2, pitching it as its most superior mannequin but for builders and on a regular basis skilled use. OpenAI had reportedly been planning to launch a brand new picture generator in early January, accelerating these plans with this week’s announcement. Its final picture mannequin launch was GPT Picture 1 in April. 

GPT Picture 1.5 arrives as picture and video mills advance past prototypes and achieve extra production-ready capabilities. Like Nano Banana Professional, ChatGPT Picture provides post-production options, offering extra granular edit controls to keep up visible consistency, like facial likeness, lighting, composition, and colour tone throughout edits. 

Most GenAI picture instruments are unhealthy at iteration, so this may be an enormous step up. Requested for a particular change, like “alter the facial features” or “make lighting colder,” fashions will usually reinterpret your complete picture, resulting in an absence of consistency. 

The replace isn’t nearly new options. ChatGPT pictures can even now be accessible by way of a devoted entry level within the ChatGPT sidebar that works “extra like a artistic studio,” Fidji Simo, OpenAI’s CEO of functions, wrote in a weblog put up Tuesday. 

“The brand new picture viewing and enhancing screens make it simpler to create pictures that match your imaginative and prescient or get inspiration from trending prompts and preset filters,” Simo wrote.

On prime of the brand new picture generator, OpenAI is introducing new methods to enhance the ChatGPT expertise with extra visible components. The plan is to make search queries show extra visuals with clear sources, which may very well be useful for duties like changing measurements or checking sports activities scores, per Simo. 

“If you’re creating, it’s best to be capable of see and form the factor you’re making. When visuals inform a narrative higher than phrases alone, ChatGPT ought to embody them,” Simo wrote. “If you want a fast reply or the following step lives in one other instrument, it needs to be proper there. As we do that, we are able to hold closing the gap between what’s in your thoughts and your capacity to deliver it to life.”