OK,
instead of speculations, here are some real measurements on warmup behavior of LTZ1000, 45°C / 12V heater, LM399H, 90°C / 15V heater, and 34465A Ohm range (similar to 34461A, as both contain LM399 and probably equivalent Ohm circuitry).
The LTZ1000s oven needs about 20sec to warm up within + 3.6ppm. The deviation at 20°C is about -900ppm, so the estimated T.C. is about +36ppm/K.
It takes about 5minutes to settle to 1ppm, the drift over the next hours is more or less related to the whole circuitry (Andreas PCB and design)
The LM399H needs about 15 sec to settle to < 4ppm, the T.C. of its reference is about -4ppm/K.
This setup on a breadboard is not so stable, I would guess, that on a PCB, it takes about 10min to settle to 1ppm.
Inside a complete instrument, it takes about 1/2h to settle to < 1ppm.. that I observed on my 34465A as well as on my 3458A.
Here's the Ohm mode shown, the volt mode should behave the same, but might be more stable after the first hour.
Rem.: The 34465As 10kOhm mode can be as stable as < +/- 0.5ppm over 12h, if stabilized for 1/2 day, and ACAL done once.
Frank