How can the same move, from the same position, be both brilliant and a mistake?

28. ... Be5 is legitimately brilliant, at least as chess.com judges these matters, because 28. ... Rc1 is defeated by 29. Rhe3. Claiming the draw with 30. ...Be5 is a mistake, because after 30. ... Rc1, 31. Rhe3 is no longer available. Black can force white to give up both rooks for black's rook and bishop leading to a QvQ endgame where black's additional pawns will prevail.
So why is 29. ... Be5 also rated brilliant? It is the exact same move from the exact same position as at move 30, and so the alternative 29. ... Rc1 would lead to the exact same win, just a move earlier. After 29. ... Be5, the win is still available, so it's not really a mistake, but it doesn't move the game closer to the win.

I get that it's a mistake. It just seems strange that it should be classed as brilliant the turn before.

The computer recognizes black's advantage and 3-fold repetition. Since ...Be5 a third time allows the game to be a draw, it calls this move a "mistake" because it blows black's advantage from completely winning to equal.

The reason the move was originally considered "Brilliant" is because it was a "good sacrifice" which kept black with the large advantage. On the board, it is the same move... but it's not exactly the same, since the mistake allows 3-fold repetition, whereas the other instances don't.

Probably a mistake

Yes