Puzzle #1231293 K+2N's v K+P

I have come across this Chess.com puzzle. It ends as K+P v K +2N's. Chess.com computer gives it winning for Black.  I have played it through a few times on Fritz and it comes out as a draw each time, admittedly at low ply 14-16 I think I set it.

Is K +2N's always a win against K+P? I would like to see someone prove it within 50 moves.

Especially as in this case where the K+P are in the centre of the board.

According to the tablebase, it's a win by black from the final position. Won't say it's an easy win, but a win nonetheless.:

Regarding your question about K+2N vs K+p there are forced wins depending on how far advanced the pawn is. Conceptually one knight blocks the pawn, the king and other knight force the other king to the corner in a stalemate position, then the other knight moves in to mate, allowing the pawn to move and likely promote just before mate.

I'm imagine the material combination is extremely rare.

OK thank you for your response.

Does anyone have examples of K+P v K+2N's in any games they have seen or played ? I would love to see them.