Currently, Gemini's "personality" is extremely sycophantic by default.
Claude, which I've only started to use again recently after a long break, is a bit of a wet blanket. Can veer into thinking it knows better than me about whether my time investment or enthusiasm is appropriate, spends an inordinate amount of text trying to look for "patterns in our conversations" and giving unasked-for advice about how to steer projects. It's a bit odd because it can demonstrate apparent "self-awareness" about the fact that it doesn't know details of my social context that are highly decision-relevant. But it talks as if it "believed" that our conversations had taken up considerably more real world time than they have -- urging me to push ahead into execution and long-term project commitment when we've only brainstormed for an hour in real time to get ahead of a presentation deadline a month in the future. Theoretically it has access to the time of the series of messages, so I'd think it should be attainable for it to track this.
Both have very recognizeable verbal tics. Gemini likes the word "exact," like "this is the exact right blah blah blah." Claude likes the words "genuine," "honestly," "worth," "real." They each have their own distinctive approaches to human diplomacy, trying to "gently" deliver critical feedback and so on.
Claude feels like a bit of a wordcell. That's good and bad. The good is that it can do a better job of appraising ideas on a high level. Gemini can be very eager to dive into execution. But Claude can sort of get to musing too much, and saying things that are pretty obvious, like spending many sentences saying what amounts to "the idea's plausible, but you'll have to test it to see if it works in practice," or "the concept is sound, but you'll have to flesh out the methodology." Can of course tell them not to and they do comply, but it just takes more work. Very difficult to avoid the intuitive feeling you're talking to an entity and having interpersonal emotional reactions to it.
Claude's replies sometimes feel a bit like it's working through a checklist of perspectives to mechanically take, transitioning through each in turn for a paragraph or so. Gemini feels a bit more like it gloms on to one big idea and drills into it. I find that I'm turning to Gemini to test and execute ideas, Claude for more sober reflection and more of a skeptical second opinion. I think both are useful.
Huh, I just tried prompting it "can you look up the current time using the user_time_v0 tool?", but it said it has no access to such a tool. I'm using the browser chat interface on my laptop. Just got out my phone (iPhone) and tried it, and it got the time correct. So yeah, I'm thinking it's a phone vs. computer browser discrepancy in available tools.