OpenAI has unveiled its future plans for AI models GPT-4.5 and GPT-5, as CEO Sam Altman shared a detailed roadmap on X. In his post, Altman recognized the complexity of OpenAI’s product lineup and expressed a desire to simplify it, stating, “We hate the model picker as much as you do and want to return to magic unified intelligence.” He mentioned that GPT-4.5, referred to internally as Orion, will be the last of OpenAI’s non-chain-of-thought models. The Verge had previously reported on some aspects of Orion.
Looking ahead, Altman emphasized that a primary goal is to unify the o-series and GPT-series models. This integration aims to create systems capable of utilizing all available tools, determining when to engage in extended thinking, and being versatile for various tasks. He noted that GPT-5 will be a system that combines much of OpenAI’s technology, including o3, indicating that o3 will no longer be released as a standalone model. OpenAI first hinted at o3 in December, with o3-mini launching in January.
In terms of user access, Altman announced that once GPT-5 is released, free ChatGPT users will enjoy unlimited chat access at a standard intelligence level. Plus, subscribers will have access to GPT-5 at a higher intelligence level, while Pro subscribers will experience an even more advanced level of intelligence. Although Altman did not provide specific release dates for GPT-4.5 and GPT-5, he suggested that they could be available in “weeks or months.”
In other news, Elon Musk recently proposed purchasing OpenAI’s nonprofit arm for $97.4 billion, but the board of directors is expected to reject this offer. OpenAI’s focus on enhancing its AI capabilities through these new models reflects its commitment to advancing technology while addressing user needs and simplifying its offerings.
