As a PRO user, I notice a clear qualitative difference between GPT-4.0 and GPT-5.x.
GPT-5.x feels significantly more cautious and heavily safety-tuned, particularly in areas like relationship advice, personal dilemmas, and life guidance.
GPT-4.0 felt more like an engaged conversational partner that offered concrete reflections and perspectives. GPT-5.x, while safer, often defaults to abstract and non-committal language, which reduces its usefulness for these scenarios.
I understand the safety considerations, but I would strongly value:
-
continued user choice to access GPT-4.0 or a similar less-filtered model
-
potentially as an opt-in / legacy / advanced option
-
especially for paying PRO users, who consciously accept the trade-offs
-
I would encourage OpenAI to actively survey users before fully deprecating GPT-4.0. For a meaningful segment of users, this model clearly provides higher value.
Thank u for reading.