With million records with unique strings, and indexed field for them, would simple find and $in holding all million variations of string perform equivalently?
If you mean thet would return the same results, yes. Performance wise, I doubt it 
But you should set up a dummy collection and try!
My computer is crap, I will not. I imagine they could just sort the contents of $in aligning it with sort of index, and just cross reference the two however fast DB architect can do such a thing.