I just watched the below video from Prusa about the recent Full Colour Spectrum advances:

Considering that neither this nor coaxial extrusion can create red from magenta + yellow (see here), it means that 6 colours are needed to access the full spectrum of hues. White and black are then needed to change the hue to the desired luminance and saturation.

Working entirely on the assumption that yellow-around-red looks exactly the same as red-around-yellow, it should be possible to only need 4 filaments being coaxialised at once. 2 of the inputs are for White and Black, and the other 2 are connected to splitters: Red, Green and Blue on one splitter and Magenta, Yellow and Cyan on the other.

Now, when I say "splitter", I don't yet have anything precise in mind. It could be a cold splitter where one material is pulled out and another is put in. It could be a hot splitter, something similar to Coaxial8or R2, where they split in the hotend. Reliability wise, I think it's the better strategy, avoiding having to carefully form the tip or need cutting hardware. This will still need heat valves, but at least the max potential power draw is cut approximately in half.