What most people would suggest is putting it in a drawer for a month at the minimum - preferably three - and then taking it out, and reading it like a reader (though with pen in hand) to see whether it all hangs together.
What most people actually do is give it one last read, correct five teeny errors (and thereby introduce three others they don't spot) and send it out. As I said
here, there will be four trivial typos, one hilarious one, and one hideous continuity error you won't spot till a reader points it out, but in 100,000 words that's okay, and they won't break a deal you would otherwise have made.
On things which are vastly important but very subjective, like voice, and is-it-too-grim (or voyeuristic, or fluffy, or whatever), I think if you've made use of beta-readers and done your best by that, and other means, to get a sense of how others read it, in the end you have to say, "Here I stand. I can do no other" and get on with it.
In other words, with book-length projects, at some point you have to forgive yourself and it a) for the tiny flaws you can still see, because nothing made on that scale can be perfect: something which is a flaw in itself is serving a bigger purpose most of which is elswhere. And b) you have to forgive it for what it can't be - gritty enough to please the grit-lovers without alienating a single fluffy bunny boiler, voicy enough for the voice-lovers without ever having to lapse very slightly in order to convey necessary information or turn an awkward plot-corner smoothly.
I know an awful lot of would-be writers who send stuff out way too early. But, equally, I know a good few would-be writers who spend five years tweaking and doubting, largely (though unacknowledgedly) from horror of that Protestant moment of "Here I stand". Those teeny tweaks really aren't going to make the difference between the book being bought and not being bought.