One small detail is, if they take folded cards into account. If 6 people were dealt in, and 4 have folded, then there are 44 unseen cards on the turn, but only 36 of those can come on the river. So if they calculate true equity, that will be somewhat different from the equity, we can calculate without knowing the folded cards.
Its unclear, if they do it this way, but it would make sense to prevent players from culluding against them by sharing information about folded cards. If for instance we are all-in on the turn, and our opponent has 12 outs, then 32/44 unknown cards will give us the winning hand, which gives us a theoretical equity of 72,7%.
But if our buddy
tells us on Messenger, that he folded two cards, that are NOT among our opponents outs, then 30/42 unknown cards gives us the winning hand, which means our euqity is now only 71,4%. So if we are offered to cash out 72,7% * 0,99 = 72,0%, we can make a long term profit accepting this offer at the expense of GG Poker.
Its also not entirely correct to assume, that folded cards are completely random. Preflop people are more likely to fold low cards than high cards, and postflop they are more likely to fold cards, that fail to connect with the board. So I would guess, that cash-out offers are based on true equity, and in that case it can explain at least small deviations from the equity, we calculate based on the assumption, that all unseen cards are equially likely to come on the turn and river.