You’ve possibly heard about this already but, if not, OpenAI has released two new models as previews for use and feedback.
These new models do much better than GPT-4o at math, logic, and reasoning tasks. They also “think” and share their thought process through a problem.
They tend to take longer to produce output than other models so if you aren’t doing a math, logic, or reasoning task, you may not want to use this.
It will be trial and error to see which models you like better for which tasks.
Here is the OpenAI post about it - https://openai.com/index/introducing-openai-o1-preview/