A beautiful position from a real opening

[OneTrickPony]

I am not sure if it's known. I came across this position analysing one of my blitz games:

[d]r1bq1rk1/p4ppp/1pnbp3/3p4/3P1B2/2PBPN2/P4PPP/R2QR1K1 w - - 0 12

And now a series of sacrifices follows:

12.Bxh7+ Kxh7 13.Ng5+ Kg6 14.Qg4 Bxf4 15.exf4 f5 16.Qg3 Kf6 17.c4! dxc4 18.d5! exd5
to arrive at:

[d]r1bq1r2/p5p1/1pn2k2/3p1pN1/2p2P2/6Q1/P4PPP/R3R1K1 w - - 0 19

Leela says 19.h4 is 60.5% for white.
SF says +0.16 at depth 65 running the analysis a few moves into the mainline.

White is probably not winning this as SF line ends in what looks like drawn Q vs 2R endgame although there are many alternatives along the way. Just wanted to share, it's a nice sequence and maybe there is something more than meets the eye.

[jefk]

https://www.365chess.com/game.php?gid=4107136

[jefk]

a rare position, Akobian also earlier (2014) played it (and lost)
https://www.365chess.com/game.php?gid=3915628
the move 12. Bxh7 isn't known apparently by Gm's; anyway after c4! maybe
17... Rh8 also is an option; although SFNNUE doesn't like it
for human players it would be hard to refute.

Anyway after 17...dxc4! d5 exd5 and then the Leela move h4,
SfNNUE says 8...g6 and then still only +0.18 for White.

NB if that's correct, then simply 12.Qe2 (which Akobian played)
instead of Bxh7+ is better (and a bit more simple).

Although later Akobian lost because most likely a blunder.

Gm's nowadays don't play at the level of SF(nnue or not)

PS chess is (at least weakly) 'solved';
it's a draw ofcourse.