anecdotally, I'm flipping back and forth between o1 and GPT-4. o1 is mildly better at editing larger code segments. I worked with it to edit a large ~2k line python file in an unusual domain.
But o1 is also incredibly verbose. It'll respond with 1-2 pages of text which often contains redundant data. GPT-4o is better in it's descriptions.
But o1 is also incredibly verbose. It'll respond with 1-2 pages of text which often contains redundant data. GPT-4o is better in it's descriptions.