You do get different inputs, but they largely share characteristics so good solutions should always work and naive ones should consistently fail.
There has been the odd puzzle where some inputs have allowed simpler solutions than others, but those have stood out.
I don't know how much they "stand out" because their frequency makes it so that the optimal global leaderboard strat is often to just try something dumb and see if you win input roulette.
if we just look at the last three puzzles: day 23 last year, for example, admitted the greedy solution but only for some inputs. greedy clearly shouldn't work (shuffling the vertices in a file that admits it causes it to fail).