This is as clear a draw as can be, as whenever black sacs the bishop for the b pawn, all the black king needs to do is move between g7 and h8, and if g7 isn't an option, its stalemate. White can never block access to h8 obviously. All endings of 7 or fewer pieces are 100% solved, and most decent engines either have the solutions to all such positions pre-loaded (called the endgame tablebases) or you can download them for the engine's use if you have space. The chess.com free analysis uses a strong engine, but I don't think it has the tablebases, and it doesn't think for nearly long enough to brute force the solution to this kind of position.
End game forced tie?
It's a draw all day in the ending. 7.g4 Bg6 8.Ne5 is a better way than 7.Be2 to challenge that Scandinavian, the knight can potentially go from e5 to c4 attacking the Qa5 and the Bg6 remains on the board doing nothing, and can be exchanged Nxg6 if it threatens to do anything, rather than the Bxf3 Bxe2 simplifying exchange that happened in the game.
Opposite-bishops endgames like these have tendencies to yield draws. Perhaps the computer does not recognise these situations and still evaluate the position by material value.
It is rather unfortunate that the Black king cannot be chased away from the g7 and h8 squares since the light-squared White bishop can never enter the a1-h8 diagonal.
So I just got done playing a Blitz game 5 minutes.