My HP34401 showed 10.00000V (+1/-2 over N=100, NLPC=10) at about 23°C (can't currently measure temperature with much confidence), which is quite astonishing given that it has not been calibrated in the last two years (if ever).
My Datron Wavetek 1271 showed between 9.999,9752 to 9.999,9766V (min, max each average over 10). Not sure, whether the variance is due to low frequency noise of the DUT or the one of the internal reference or some draft across the contacts. My home 'lab' is not suitable for sub-ppm measurements. This result too is IMHO quite remarkable, given that the instrument was calibrated in 1993 (or at least nobody since bothered to update the calibration time stamp in the NVRAM).
I think this goes to show that with some luck old instruments can be very stable and much more accurate then the nominal specification suggests.