New user: ChatGPT value as a general chat & research tool
retained text
Relative newbie to AI tools, probably a total of about 6 months. Main use cases are general chats, research, brainstorming, document/spreadsheet analysis and some light agentic coding & technical/architecture reviews (for work and personal). I started using Gemini, Claude and Opencode (various models) initially, and in the last month or so started using ChatGPT quite a bit. I would often take the prompt from one and run it in another one, setting comparable models & effort levels, and compare the results. My methodology was neither scientific nor rigorous, however the last month of nearly daily use has highlighted for me that ***ChatGPT generally seems far superior for all the non-coding tasks I mostly do***. The best way I can explain it is that Claudes's answers are often sound, but somewhat lazy, not very rigorous, never dig in past first layer or two (regardless of model or effort). Also default output layout and structure always leave much to be desired. I have to really poke and prod it to get a truly useful output. On the other hand, ChaptGPT seems to do better at understanding intent, goes deeper, and provides more logical and structured output that easy to understand and digest. I post this as I keep hearing all the doom and gloom over on the codex sub, with lots of people jumping ship. Much of this is justified and self-inflicted by OpenAI, but I want to stay on topic of ChatGPTs vs Claude usefulness as a non-coding tool. Do you have the same or different experience, particularly with more recent models? Thanks!
document fetched 2026-10-11 00:20 · canonical https://reddit.com/r/ChatGPT/comments/1x2ijc0/new_user_chatgpt_value_as_a_general_chat_research