Okay, who is a mentor? A mentor is one person, specific, who has done something close enough to what you're trying to do that when they tell you something, you do it first and understand later. That's the clearest whole definition that should matter to you.
How many of them do you need? Just one. I'll explain.
See, the conventional wisdom suggests otherwise. Your gravitational pull will be to go and ask five people, take their opinions and advices. You'll likely get partially-overlapping opinions. The funny thing is after you synthesise all the conversations, you'll direct yourself to a lot of confirmation-bias and the synthesis will almost always land on whatever you'd quietly wanted to do anyway. Those five people? They are not accountable for what you do next. You aren't accountable either, because you were following advice which is averaged… the kind you get from a committee.
You also don't realise that you're doing administrative work instead of real work. Say you have the five mentors and you're setting up calls or meet-ups, then taking the calls and meeting them. If you're doing this seriously, the entire administrative work will take the better part of a working month. As a result, the work you should be doing, the thing that was supposed to make you worth mentoring in the first place gets pushed behind. You're busy getting mentored and that looks like progress.
At the end of the day, conversations are interesting and the actual work is hard. Of course the conversations will win.
One mentor doesn't give you that kind of exit.
Okay. One mentor.
It's likely that the one you pick will occasionally be wrong. That's the point.
Pick one. Follow them sometimes when you don't agree. Some advice will land and some won't. When it lands you'll learn something. When it doesn't you'll learn something sharper: exactly what their blind spot is, exactly where your instinct was better, exactly what you'll do differently next time. People who've followed one person for a few years know precisely where that person is strong and where they aren't. People with five mentors have spent three years hearing opinions and have no idea whom to trust on what.
Now, the question is: Who do you pick? And I don't have a clean answer for it.
Someone who'd done something close to my actual problem is a great starting point. Not the same problem necessarily but even if it's close enough. It could be someone specific where you need specificity, and who didn't hedge when you needed a call. Someone who'd pay attention to your actual situation and not to a general version of it. When that's there you can feel it. When it isn't you get answers that would work for a general case study, which is what most startup mentorship actually is.
That's most of it. You don't need that from five people. You need it from one, over enough time that they've watched you be wrong and seen what you did next.
The advice industry has made the rigorous-sounding multiple mentors version sell better, and it sells better to people who'd rather keep discussing than go and do the actual work.
At the end of the day, you'll like having hours of conversations and feeling sharper than you were an hour ago but they also leave you exactly where you were an hour ago.