I was mostly using 4o for academic searches and planning. It was the best model for me. Based on the context I was giving and questions I was asking, 4o was the most the consistent model.

It used to get things wrong for sure but it was predictable. Also I liked the tone like everyone else. I stopped using ChatGPT after they removed 4o. Recently, I have started using the newer GPT-5 models (got free one month). Better than before but not quite. Acts way over smart haha