Queen vs 2 Rooks on Larger Board Sizes

On 8x8, the Queen is typically worth more than the nominal "9", at 9,5, and tends to outperform the Rooks in early to early-mid game, and can even steal draws when the rooks are winning late, making the trade not always clear, especially if the trade costs tempo.
.

In AlphaZero-derived piece values, the queen's 9,5 remains consistent between classical and "no castle" chess, while the rooks go from 5 apiece in no castling to 5,63 in castling-enabling classical chess, suggesting castle rights have a separate value of 1,26 from the rooks' base value.
If you break down the Queen's value into its likely components: 5 (rook) + 3,25 (bishop) + 0,5 (built-in bishop pair, since it can hit both colors) + x = 9,5, That leaves 0,75 left to be x, which likely comes from the new orthogonal connections ferz and wazir components of the rook and bishop make with each other.

.
The next question is this, if all of a rook's orthogonally adjacent slide scales well with board size, but bishop + bishop pair scale less, and the orthogonally adjacent connections right next to the queen don't maintain their size relative to the board size, and she suffers from leveling effect and major piece redundancy effect with larger armies, how fast does the queen's value decay compared to the rooks on larger board sizes? Also, let's not forget that the value of castling is a massive 1,26 pawns at engine depth with it only providing double tempo swing for the king, what value would it provide if the king moves, 3, 4, or even more spots, relative to its normal speed of 1?
.
Curious to know everyone's thoughts.