I’m not sure if I’m entirely right. If I’m wrong, I’m happy to be pointed out.
I think the difference is when you use o1-model, it doesn’t have access to the long-term memory. While with the o1-reasoning tool, you can have the memory held by ChatGPT-4o, and have it use o1’s reasoning ability, which opens another door, in my opinion. At least, with the way I use them.
And, I think the hybrid mode is really nice. When you use purely o1-model (upper-left corner), you can’t discuss in a very casual way after it explains something deep to you since it’s going to count as o1 (plus users: 50 times per week). With this o1 reasoning tool, I think I can keep using 4o and it understands the context so far. When you need to ask something difficult, you don’t have to copy them all and paste to o1 model, and hope it gets them all. You can just use o1 reasoning ability to talk with your 4o model.