Would be interesting to see the real comparison values between the calibrators.
Yes, that would be very interesting, but seems very unlikely to happen.
The Fluke 5720/5730 and the Wavetek 4808 are the very best multifunctions calibrators and seem to stay well below their specs over time.
But with difficult DUTs like HP3458A (or the even more challenging Fluke 8508A) they are still just not good enough. Therefore you have to have a "golden" calibrator which needs to be characterized at least within every 90 days. Afaik Fluke uses a 45day Intervall and is able to achieve 24h specs over time.
What I can do and have done is of course not traceable and therefore I do not know real absolut values. But when I compare my own gear, I read values against my own Standards and my best DMMs. But with this I still have good confidence, because most of my long scale DMMs agree within around +-1 ppm when checked against my very stable 10V DC reference.