OpenAI CEO Sam Altman has announced the upcoming release of two new language models—O3 and O4-mini—with the highly anticipated GPT-5 to follow a few months later. The announcement was made via a post on X (formerly Twitter) and marks a shift from the company’s previously suggested plan to go directly to GPT-5.

New Models to Precede GPT-5 Launch

The rollout of O3 and O4-mini will serve as a stepping stone toward GPT-5, allowing OpenAI to gradually scale its systems and test improvements before the main release.

“There are a bunch of reasons for this,” Altman wrote, “but the most exciting one is that we are going to be able to make GPT-5 much better than we originally thought.”

Improvements and Infrastructure Readiness

While Altman did not provide technical specifications for the O3 and O4-mini models, he hinted that O3 would deliver significant improvements. “I think people will be happy,” he added, suggesting a better performance than earlier iterations.

He also acknowledged challenges in integrating new features and ensuring infrastructure capacity to meet what he called “unprecedented demand” expected with the launch of GPT-5.

GPT-5 Expected in Second Half of 2025

Although no exact date was shared, Altman mentioned GPT-5 will be released “in a few months,” pointing to a likely launch in the second half of 2025. The company aims to use the releases of O3 and O4-mini to collect feedback and refine systems before GPT-5’s deployment.

Read More: OPPO Reno 14 Pro Leak: Flat 120Hz Display, Triple Cameras & New ‘Magic Cube’ Button Expected