I've been trying to have the free GPT-3.5 explain me things, but the results were a bit mixed (I caught it making a few things up earlier). But I might give GPT-4 a try, thanks!
GPT-4 is significantly better than GPT-3.5-turbo. Do not use GPT-3.5-turbo outputs as your benchmark for the utility of the more advanced models! :) Especially when used for LLM Agents, or any kind of harder question.