Part 2/9:
One of GPT-5's notable innovations is its ability to auto-select the appropriate model for your request. For simple questions, it employs a lightweight, fast model, whereas complex prompts trigger a heavier reasoning model. This dynamic is designed for cost efficiency, but it can sometimes lead to responses that feel underpowered or superficial, especially when more nuanced answers are needed compared to GPT-4.
Two straightforward solutions can improve results:
Prompt the model explicitly to think deeper by adding instructions like, "Think longer about this," which nudges GPT-5 to invest more computational effort.
Manually select a dedicated "thinking" model from the top menu to guarantee a more detailed response.