I think the industry leading part is "having 8 channels at 6.25GS/s", id say that's pretty impressive for the category this scope is aimed at.
The datasheet for the Tektronix 5 series does clearly state:
8 bits @ 6.25 GS/s
12 bits @ 3.125 GS/s
13 bits @ 1.25 GS/s (High Res)
14 bits @ 625 MS/s (High Res)
15 bits @ 312.5 MS/s (High Res)
16 bits @ ≤125 MS/s (High Res)
But yeah it is a bit misleading calling it a 12bit scope (And it does have a 12bit converter in it) and then also slapping on the 6.25GS/s number.
The Tektronix 5 series datasheet does appear to downplay the RMS noise performance of it (If not, then the AFE in this scope is quite a noisy one):
All of this claims to be in High Res mode. Keep in mind that the values for < 10mV/div 50Ohm on the 2GHz model start dropping off in bandwidth down to only 175MHz (1mV/div) even tho the column is labeled 1GHz
Compared to the Keysight Infiniium 9000 series that is about 10 years old (the 9014A is the 1GHz model):
Where no special high res mode is used on this scope, while only having a 8bit ADC (I did check on my scope, the table matches when running in 50Ohm, Normal mode 20GS/s). The numbers look pretty similar to the Tek with its 12bit ADC.
However the Infiniium 9000 does also have a high resolution mode that can be turned on on top of this, but when it is turned all the way up to 12bit the sample rate drops off to only 2.5GS/s and the channels bandwidth drops to 550MHz but it also results in a RMS(AC) noise of
99uV on the 10mV/div range. Dialing it back to 11bit in order to get us to a more Tek comparable 5GS/s 1.12 GHz of bandwidth and a RMS noise of
150uV. So significantly better than the 12bit Tek
Infact looking up the LeCroy HDO6000 (Another true 12bit ADC scope) the RMS noise spec for 1GHz 10mV/div is 155uV. This is 5uV worse than my Keysight 9000 that is pretending to be 12bit on its old 8bit ADC.
As you see having a 12bit ADC might not be as big of a advantage as it sounds.