This is the famous Behting study, first published in the magazine Bohemia in 1906, which is often used as a demonstration that computers can’t solve all positions.
White to move and draw with a very elegant move.
1. Kc6!! g1=Q 2. Nxh4 Qh1+ 3. Nhf3
Question is, is this study still beating engines?
Clarification: Suggesting best move, doesn't mean engine understands(or finds) the solution. Without evaluating it 0.00 or =/=, it doesn't mean it's solved. Actually, best move with -4.** means engine believes that best possible continuation for White is this move, with winning advantage for Black.
Imagine you ask a question to someone about a chess position in which White has a move which is the best and draws. You ask, which move is the best and why, and you get an answer like this: "Best chance is this move, but even if it's White's best chance, it looks losing". And you know that move draws. Would you accept this answer as correct answer? Definitely no.
I recommend to read:
- A very instructive article about interpreting evaluations of chess engines: How to Read Engine Evaluations?
- This post: Computer evaluations: How trustworthy are they?
Notes about edits: As the question turned into a discussion about whether the position solved by engines or not, I revised the question accordingly.
This question is totally about facts, and not opinion based. Some of the answers are opinion-based, but it's not the problem of the question. It's terribly wrong to close this question, instead of letting people discuss it.