Exactly that.
The advantage (if there is one) is that the channel only handling the mid/tweeter horns will be operating at lower distortion levels as the current hungry, distortion inducing bass is handled by a separate amplifier channel. As the mid/tweet amp will basically be operating soooo far inside it's output/distortion envelope you may find the system appears to play louder due to a cleaner HF signal.
Pondering out loud:
I suppose you could debate the above, because each of the four channels will be receiving and amplifying a full range signal. This after all is not an active setup. That said, the lower half of this signal hitting the midrange high-pass filter is probably a somewhat less reactive load than the bass driver, so I'd wager the mid/treble amp is still getting an easier ride.
I suppose you could also debate whether you should use a PA5 per channel (vertical bi-amping) or one for the mid/treble of both speakers and one for the bass - horizontal bi-amping. I suspect given the vanishingly low cross-talk measured in the PA5, the vastly more complex NAD C658 is the limiting factor, thus rendering this issue moot.
Either way, you don't passive bi-amp to gain a power increase.
Indeed, I don't know if I'd bother given that the whole thing doubles the price for the minimal gains. I only did it in this instance because it was so cheap to do so!