I missed most of the 478A /432A discussion, I went to bed early. Most has been said already but in summary:
It's the head (475A) that is important not the meter. The meter just need to balance the thermistor bridge, an external digital meter makes the measurement in the metrology set-up
No, there is nothing else that makes an easy, portable RF power transfer standard. All you have to ship is the head. The 432x does not have to be calibrated and the DVM/DMM you use with it is just local calibration.
The new N432A was built because hey culd not get parts to keep making the 432A. They could have just done a bridge balancing box for use with an external DMM but thet would not have required calibration so less through life revenue for Keysight. To be fair it does do he calculations for you so is easier to use.
The calibration reference for them is another thermal device, probably a fluid based calorimeter system.
The military / aviation "you must use an ACME 123ABC meter" is different. Often it is just a contempory standard meter that could be replaced by something else. The trouble is you don't know for sure unless the manual says "or similar". Ive said before that failure to use the specified test equipment on a angle of attack vane was one of the factors in the Lion Air 737MAX crash. I've personally seen a case where my favorite DMM Fluke 8060A was called up and the workshop was using an 89IV this gave different results. Trouble was it was fine on the bench with either but in the "corner-case" (low, slow, hot / high, fast cold etc) where it really mattered the instrument would misread if set with a 89IV. The voltage being measured was a kHz range badly distorted sinewave. The only other meters that worked were thermal ones like HP3404 or Fluke 8920
And there we are back at thermal