Does anyone know which of the methods is most efficient, and which, when these technologies are fully mature, is likely to then be the most efficient? Whatever the answer, that's the standard that I wish to see prevail.
None of these will be particularly efficient unless the device is sitting directly on the transmitter with both the transmitting and receiving coils lined up.
But because the TX coil will likely be much larger than the RX coil (needs to be if you want to put multiple devices on the same charger without having to worry about lining them up perfectly), only a fraction of transmitted power will ever cross a receiving loop and the rest will be lost even at point-blank range. Once you start putting distance between the pad and device, you quickly get in the far-field where free-space dispersion will increase losses proportionally with d^2 which makes wireless power at "long" ranges woefully inefficient even on a good day unless you use highly directional antennas but that would ruin the convenience argument for wireless charging in the first place.
I'm betting one of the inductive schemes will win since it is much simpler and cheaper: simply drive one or more simple coils with simple inverters (which coils/inverters are being driven may change depending on where loads are located on the pad to minimize waste) and use a simple coil with rectifier bridge on devices to power them.
With resonant, you have the same fundamental building blocks as inductive with the added burden of managing each device's capacitive/inductive load to keep the near-field's overall LC load and resonant frequency roughly constant, which calls for coils and capacitor banks with multiple taps and a management layer to negotiate LC configuration across all devices within the charging field.
Resonant circuits are nice for picking weak signal sources over narrow bands but when the primary application is practically direct contact with the charging pad, its sensitivity (and associated complexity) is unnecessary.