was 25. Bxe4 really a mistake for white?

en.lichess.org/aVyp31Lc/black#48

Curious why the computer thinks 25. Bxe4 was a mistake for white?

Followed it up with several stockfish variants from positions between 25 and 29 and given best play by black/white it always lead to a drawn position.

Is this just a case of the move horizon compromising the analysis or am I missing something?

If I'm not missing something, was 24. Ne4 actually a mistake in that case?

you weaken lightsqaures and you give the bishoppair and give the square on d5

I have a feeling that the computer dislikes 25.Bxe4 because the knight landing on d5 is quite uncomfortable for white. I understand the logic behind 25.Bxe4, as the knight looks does look quite strong on e4, but after a move like 25.R3c2 the knight isn't threatening anything directly, whereas the knight landing on d5 threatens the f4 pawn, the DSB, and the rook at the same time.