OpenAI's New Plan: o3/o4-Mini Soon, Better GPT-5 Later!

Heads up from San Francisco! OpenAI's CEO, Sam Altman, just shared a bit of a switch-up in their plans for rolling out new AI models. Get this: instead of jumping straight to the big one, they're actually going to release two models called "o3" and "o4-mini" first. We should see those pretty soon, probably in just a couple of weeks! The much-talked-about GPT-5 is still coming, but it'll be a few months later than maybe expected.

Altman spilled the beans in a tweet on Friday. He said, "Change of plans: we are going to release o3 and o4-mini after all, probably in a couple of weeks, and then do GPT-5 in a few months."

So, why the wait for GPT-5? Well, Altman gave a few reasons, but the really exciting one is that they think they can make GPT-5 way better than they first planned! He mentioned, "the most exciting one is that we are going to be able to make GPT-5 much better than we originally thought." Sounds like they've had some cool ideas or breakthroughs they want to bake in before releasing it.

Besides aiming for a more powerful GPT-5, Altman also acknowledged the practical challenges of development. He admitted it was "harder than we thought it was going to be to smoothly integrate everything."

Plus, they're getting ready for a ton of people wanting to use the new models. Altman added, "We want to make sure we have enough capacity to support what we expect to be unprecedented demand." Basically, they want to make sure their servers don't crash when everyone rushes to try it out!

So, the good news is we get to play with o3 and o4-mini soon. These could be smaller, more focused versions of their larger models, possibly designed for specific tasks or to be more efficient. Meanwhile, the OpenAI team gets more time to make GPT-5 extra awesome. It just shows how fast things change in AI, right? This revised plan prioritizes both immediate access to new technology and a more substantial leap forward with GPT-5.