Recently, Sam Altman, the CEO of OpenAI, shared a development roadmap on the X platform, outlining plans for the upcoming GPT-4.5 and GPT-5 artificial intelligence models.

Altman acknowledged in the post that OpenAI's product line has become somewhat complicated. He emphasized the company's commitment to simplifying its product offerings. He mentioned that both OpenAI and its users are not fond of the "model selector" and aim to return to a more unified "magical" experience.

It was revealed that the upcoming GPT-4.5, internally referred to as Orion, will be the "last non-chained reasoning model" released by OpenAI.

Following GPT-4.5, OpenAI's primary objective is to unify the O-series and GPT-series models by developing systems capable of utilizing all available tools and knowing when to engage in deep thought processes, thus enabling broad application across various tasks.

For ChatGPT and OpenAI's API, the company intends to build GPT-5 as an integrated system incorporating multiple technologies, including o3. Altman also stated that OpenAI will no longer release the o3 model independently, having already disclosed details about o3 in December and launching o3-mini in January.

When GPT-5 is released, free ChatGPT users will have access to "unlimited chat" under the "standard intelligence settings." Plus subscribers will be able to utilize the "higher intelligence level" of GPT-5, while Pro subscribers will gain access to "advanced levels of intelligence."

Regarding the exact release dates for GPT-4.5 and GPT-5, Altman did not provide specific timelines, estimating only that they would arrive within "weeks to months."