Considering that all have high distances comparing to g25 it looks like some Gedmatch oracle (single population mode) on ancient samples...
BTW why some excition:) Everybody with G25 coordinates could produce such fit list (without hidden pops) with ancients in nMonte Runner or in R.