o1-pro is good, but not perfect. It's got a different prompt style, I found this post quite helpful: https://www.latent.space/p/o1-skill-issue

The main complaint for me is that by not seeing the output of the reasoning model the experience is worse. The summary model overtop the reasoning output often gets very abstract. You know how like when Claude will write code and then it gives you a couple paragraphs of text after explaining what the code does? o1 pro will often just output that paragraph of text. Which is annoying since what I really want is the raw code to copy and paste!

For that reason you have to be very explicit about what you want o1 to give you, otherwise it's often just a summary of work done behind the scenes.

I've also found that o1 can get argumentative once the context gets long enough that it's "reasoned" itself into a conclusion. Getting it to admit something is wrong is often hard.

I'm finding it useful, but it's a very different style of interaction. And tbh, it's kind of annoying they forced it into a chat bot style UI. I'm excited for all the other players to release their reasoning models, I think there's a LOT of room to innovate on the actual interaction style with these kinds of models!