10uV/div is of course benchmark. Question how is the noise floor / which bandwidth do you have to set to take measurements?
My scope has only 2mV/div but has up to factor 50 of zoom (which is possible due to 16 bit ADC) resulting in 40uV/div.
At 100 kHz the noise floor is around 100uVpp.
Tektronix was quantitative and conservative about their noise specifications so you can just look it up. The 7A22 is better than 16uVrms over its entire 1 MHz bandwidth which is about 80uVpp. Assuming that the 1/f corner is significantly lower (I suspect it is about 10kHz), over a 100kHz bandwidth, that should be 25uVpp so about 5 times better than your oscilloscope.
There are a couple of things worth noting. The 7A22 has a fully differential input so its noise is automatically 3dB higher than a single ended design. Its input stage also has a cascode to support high common mode voltage with high common mode rejection and this also increases input noise. Yet with all of that, it still has lower noise simply because its low bandwidth and 47pF inputs allow the use of larger geometry low noise devices.
To get some idea of how much worst case (1) added noise comes from a differential input with cascode for high common mode range makes, compare the 100 MHz differential 7A13 with a standard singled ended 100 MHz vertical amplifier; the former has 400uVrms noise specified although I measured both of mine at about 200uVrms while I measured the 7A18A at about 20uVrms so a ten times difference. On the other hand, while the 7A13 is very noisy, its common mode rejection can make measurements that are impossible for the 7A18 because of added common mode noise.
The differential input is actually an advantage despite doubling the input noise because such low levels of noise would be compromised by the ground loop produced using a single ended input. *All* of the low input noise and high sensitivity amplifiers that Tektronix made, and there were several, have differential inputs for this reason. Without the differential inputs, the low input noise and high sensitivity are useless. There is a reason the 100 MHz Tektronix P6046 low voltage differential probe was such a popular product; its differential inputs allow measurements which are otherwise impossible.
Could a modern design do better? I doubt it except maybe at the lowest frequencies where 1/f noise dominates. Modern JFETs are only slightly better than what Tektronix used.
(1) The 7A22 is slow enough at 1 MHz that its first stage has gain unlike the 100 MHz 7A13 so the added noise is less. Faster oscilloscope amplifiers cannot get away with that.