OpenAI has generally been tight-lipped about the next generation of its frontier artificial intelligence (AI) model, GPT-5. However, the company has recently shared some hints at what users can expect once the model is available. An OpenAI executive has claimed that the GPT-5 large language model (LLM) could witness the merger of the GPT and o-series models. This means the upcoming LLM will feature both advanced reasoning capability as well as multimodality support. Notably, while there are no confirmed dates, the GPT-5 model could be released in July or August.
GPT-5 Might Be OpenAI’s Most Feature-Rich AI Model
Romain Huet, Head of Developer Experience at OpenAI, recently hosted a session at the Viva Technology conference in Paris. A clip from the session was shared by Tibor Blaho, lead engineer at AIPRM, on X (formerly known as Twitter). Huet highlighted that with GPT-5, the San Francisco-based AI firm is planning to build a “net new great frontier model.”
“So, the breakthrough of reasoning in the O-series and the breakthroughs in multi-modality in the GPT-series will be unified, and that will be GPT-5,” the OpenAI executive added.
This information is not entirely new, as the OpenAI CEO, Sam Altman, had said in an X post, “I would love for us to be able to merge the GPT series and the o series in 2025.” Altman has also spoken about discontinuing the o-series in the past. However, so far, it was not known if the company was planning the unification with the GPT-5 model or other AI models after this.
GPT-5 AI model might also witness an integration with the Operator AI agent. In a Reddit post, the company’s Vice President of Research, Jerry Tworek, mentioned that the focus with the upcoming frontier LLM is to make it so feature-rich that users don’t have to constantly rely on model switching.
The executive hinted that apart from making the new AI model perform better than all the previous models in every metric, it can also be integrated with the Operator AI agent. For the unfamiliar, the Operator is a Computer-Using Agent (CUA) which can perform various tasks on the user’s device and online, autonomously. It was first unveiled in January, and then expanded to Pro subscribers in February.
While a concrete release date is not known, Altman has recently hinted in a live stream that GPT-5 could arrive in summer. Since only two months are left in summer (based on the US climate), the AI model is expected to arrive either in July or August.
