Most analysis tools reserve their top label, Chess.com calls it Brilliant, Rookie calls it Amazing, for a narrow, specific case: a move that looks risky or even losing on the surface, usually because it gives up material, but is actually the single best move available and works for a real, calculable reason.
It is not just the engine's top choice
That distinction matters. An engine's literal best move happens on nearly every turn of every game; it would be meaningless to celebrate it every time. The special label is reserved for moves that are best and counter-intuitive at the same time, typically a sacrifice that a reasonable player would expect to be bad, but that turns out to be objectively correct, often the only move that holds the position together or forces a win.
A queen sacrifice that leads to forced mate three moves later is the classic example. So is a piece sacrifice that looks reckless but is actually the only way to stop an opponent's attack. What they have in common is the gap between how the move looks and how it actually performs. Small gap, ordinary good move. Large gap, brilliant.
The classic shape: Qg8+!! looks like giving away the queen for nothing. It's forced (Rxg8 is the only legal reply), and then Nf7# is mate. Handing over the most valuable piece on the board to force checkmate two moves later is exactly the kind of gap that earns the label.
Why the bar matters
If every strong move got flagged as brilliant, the label would stop meaning anything. Keeping it rare is what makes it worth celebrating when it actually happens, both for your own sense of progress and for the tools that are trying to give you an honest picture of your game.
Rookie grades every move you play against this same standard and reacts differently to an Amazing-level find than to an ordinary good move. Import a game for free to see if you have already played one without realizing it.