OpenAI CEO Sam Altman has detailed the company’s plans for its upcoming GPT-4.5 and GPT-5 AI models.
Image source: OpenAI
In a post on X, Altman acknowledged that OpenAI’s product line has become complex and said the company wants to work on simplifying its offerings. “We hate model matching as much as you do and want to get back to magical unified [artificial] intelligence,” Altman wrote.
The company plans to release GPT-4.5, which Altman says is internally called Orion, and will be “OpenAI’s last non-reasoning model.” OpenAI trained the Orion model in October last year, and it was later revealed that the company faced high costs and a lack of data in training it.
ChatGPT is currently free and offers three AI models to choose from
With ChatGPT and the OpenAI API, the company plans to “release GPT-5 as a system that combines many of our technologies, including o3 (ed. note: o3 is OpenAI’s most advanced AI model to date),” Altman says, adding that “we will no longer ship o3 as a standalone model.” OpenAI unveiled o3 in December and launched o3-mini in January.
Altman said that after GPT-5 launches, free ChatGPT users will have “unlimited access to GPT-5 at its default intelligence settings,” but with some restrictions to protect against abuse. Meanwhile, paid Plus subscribers will be able to use GPT-5 at a “higher level of intelligence,” and Pro subscribers will get “even higher levels of intelligence.”
Altman did not specify when exactly GPT-4.5 and GPT-5 would be released, but noted that it would take “weeks/months.”