Lab Notes · 1

Same Question, Different Answers: ChatGPT vs. Claude

Jun 2, 2026 · AI Note Lab

Same question, two very different personalities
Same question, two very different personalities

While writing reviews I kept feeling that "these two AIs have different personalities" — but I wanted a record, not just an impression. So I pasted the same 10 questions, spanning different types, into both AIs and compared the answers side by side.

Method

Results: 4 differences that stood out

1. Structure — bullet lists vs. paragraphs

ChatGPT produced "report-style" answers — headings, bullets, emoji — for 8 out of 10 questions. Claude leaned toward paragraph-based, essay-like explanations for the same questions. ChatGPT was easier to skim; Claude was easier to actually read.

2. Judgment questions — laying out options vs. taking a side

On "which of these two should I pick?" questions, ChatGPT tended to build a pros-and-cons table and end with "it depends on your situation," while Claude tended to walk through the conditions and then commit: "if it were me, I'd pick this one." When I was stuck in decision paralysis, the latter was a relief.

3. Writing tasks — nailing the format vs. capturing the nuance

The farewell email experiment was fun. ChatGPT wrote a flawlessly formal "model answer," while Claude wrote something with an actual emotional thread — you could feel the "I'll miss this place" in it. The one I could have sent as-is without it feeling off was Claude's.

4. Facing the unknown — answering anyway vs. stating the caveat

On a deliberately ambiguous question ("is this allowed under my company's policy?"), ChatGPT gave a long answer in generalities, while Claude more often led with the limitation: "I can't know your company's policy, so speaking in general terms…"

What I learned

This experiment used free accounts with default settings. Custom instructions (system prompts) can dramatically change both AIs' styles, so read the differences here as their "default personalities."
← All Lab Notes