Sam in a recent event in Tokyo. In the same day they launched Deepresearch. Shared some things about the upcoming models.
He said that they are planning to come up with a unified orchestrator based model . So it reasons when it needs to talks to voice when it needs to generates images and doesn't think when it's not needed.
And that could be the GPT-5.
Supposedly if it will be powered by a model that is more powerful than o4 which they said will be 50th best coder in the world. Than that will be awesome and o3 will be like the cat here. :)
コメント