I'm really interested to know how seasoned ChatGPT users pick between the various models like GPT-3.5, GPT-4, GPT-4 Turbo, and others. If you frequently use ChatGPT, I'd love to hear which model you prefer as your main tool and what particular tasks you've found each model excels at. Have you noticed any unique strengths or weaknesses in any of them?
2 Answers
I've found GPT-4o to be my go-to for coding tasks. It's really good at logical problem-solving, even though it can be a bit stubborn at times. I had a situation where it confidently claimed a function existed, but it turned out to be from a Reddit comment, and it misinterpreted it. If you know how to handle it, you can get some solid results—but be ready for the occasional hallucination!
If you don’t code much, I think GPT-4 is the best choice for general conversations, especially in humanities or social sciences. GPT-3.5 is fantastic for deep discussions, and switching between the two models can enhance the experience. You can seamlessly compare insights and get a more rounded view—kind of like having a tennis match of ideas between the two!
I can’t afford the pro version right now, though.
I thought the models couldn't keep context? Does switching just refer to the last response?
Yeah, I feel like the mini versions tend to be logical but sometimes don’t accept when they’re wrong, even with proof!