I'm feeling a bit overwhelmed by the different AI models available. I usually just use 4o since it's the default and I can work with it pretty easily. However, I can't shake the feeling that I'm missing out on something better. As a Plus subscriber, I've tried 4.5 quite a bit, but I often hit the limits and end up going back to 4o. I've also experimented with deep research models and occasionally switched to one of the other o# models. The main issue is that I don't really understand the differences between them aside from the basic descriptions, and when I ask GPT, it usually just suggests sticking with 4o. Is there more to this decision than just reverting to the default?
2 Answers
I can relate! I felt like we had access to a stronger model that resembled the current 4.5, but then it scaled back to 4o, which can feel pretty inadequate sometimes. I switched to 4.5 for a bit, but it only allows a few prompts before locking me out for a week. I hope we eventually get back to the features of 4.5 because 4o feels pretty weak in comparison.
Honestly, I think o3 is the best model for most tasks, followed closely by o4-mini. But I get it—rate limits can be a pain, and they can keep you from using them freely.
I gave o3 a try after hearing this, and honestly, I found it underwhelming. Sure, it generates solid code, but it seems to focus too narrowly on specific parts of a project, missing the big picture and causing side effects. For the way I work, I'm sticking to 4o/4.5.
What makes o3 better than the others, though? Is it more suited for specific types of work, or what do you mean by 'the best model'?