OPENAI cancels its O3 AI model in favor of a new generation “unified” version

Openai actually canceled the exit from O3which was to be the next major version of the company’s AI model, in favor of what CEO Sam Altman calls a “simplified” product offer.

In a Publish on x WednesdayAltman said that in the coming months, Openai will publish a model called GPT-5 which “integrates many [OpenAI’s] Technology ”, including O3, in its chatbot platform supplied by AI Cat and API. Following this decision of the roadmap, Openai no longer plans to release O3 as an autonomous model.

The company initially declared in December that it planned to launch O3 at the start of this year.

“We want to do a better work to share our planned roadmap and a much better job simplifying our product offers,” wrote Altman in the position. “We want AI to work simply for you; We realize how complicated our model and product offers. We hate the model picker [in ChatGPT] As much as you do and want to return to Magic Unified Intelligence. »»

Altman also announced that Optai planned to offer unlimited chat access to GPT-5 to the “standard intelligence framework”, subject to “abuse thresholds”, once the model is generally available. (Altman refused to provide more details on what this parameter – and these abuse thresholds – imply.) Chatgpt plus subscribers will be able to execute GPT -5 at a “higher level of intelligence”, said Altman, While Chatgpt Pro subscribers capable of executing GPT-5 at an “even higher level of intelligence”.

“These models will incorporate voice, canvas, research, deep research, etc.,” said Altman, referring to a range of Openai features has been launched in Chatgpt in recent months. “”[A] The best goal for us is to unify [our] The models by creating systems that can use all our tools, know when to think for a long period or not, and generally useful for a very wide range of tasks. »»

Before the launches of the GPT-5, Openai in the coming weeks will publish the GPT-4.5, named “Orion”, which, according to Altman, will be the last “non-chain model of that of the company”. Unlike other so -called O3 and Openai reasoning models, models without chain of thoughts tend to be less reliable in fields such as mathematics and physics.