A growing number of ChatGPT users trying Claude for the first time are pointing to the same thing: Claude tells you when it does not know something.
That sounds like a minor detail, but in practice it changes how much you can trust the output. ChatGPT has a well-documented tendency to answer confidently even when it is fabricating details (what the industry calls "hallucination"). Claude takes a different approach. When it lacks sufficient information, it says so directly rather than generating a plausible-sounding guess.
For anyone using AI tools for research, fact-checking, or client-facing work, this distinction matters more than benchmark scores or context window size. A wrong answer delivered with confidence is worse than no answer at all, because you are less likely to double-check it.
Anthropic has been explicit that they train Claude to express uncertainty, and it shows. The practical tradeoff is that Claude sometimes refuses to attempt things it could probably handle, which can feel overly cautious. But for most professional use cases, that is the right failure mode. You would rather have an assistant that says "I am not sure about this" than one that invents a citation that does not exist.
None of this means Claude never makes mistakes. It does. But the pattern of admitting uncertainty rather than bluffing through it builds a different kind of trust over repeated use.