Openai has changed its plans and will finally launch O3 and O4-mini, after announcing its cancellation in mid-February to bet on the unification of the two main families of models. The reasoning models, distributed under the ‘O’ family, will be expanded with O3 and O4-mini, which OpenAi will launch “probably in a couple of weeks”, as confirmed by the executive director, Sam Altman.
Initially, Altman himself announced in February that O3 would not be launched as an independent model, but part of its technology would be integrated into GPT-5. The reason is that integration was going to be more difficult than they had initially thought. “We will be able to make GPT-5 much better than we originally thought,” Altman said in a shared statement in X (former Twitter), which has also said that “in a few months will arrive.”
The shared roadmap two months ago raised the unification of the two main series of language models, GPT I, with the aim of offering “systems that can use all our tools, know when to think for a long time or not and, in general, be useful for a wide range of tasks.”