There are several issues being conflated here, though. Cell group balancing needs a slow charge rate - the balancers in the BMS just cannot handle high shunt currents.
The pack can go without cell group balancing for a few dozen charges, as in general, the design of the pack is tolerant of, and self-compensates to some degree, for the small differences in capacity and internal cell resistance between individual cells (the main reason for cell groups to drift out of balance).
The BMS seems to only actively balance cell groups when the opportunity arises, which seems to be to top balance at the end of an AC charge, when the BMS can actively control the chargers to reduce charge current to within the dissipation limit of the shunt resistors. At all other times, the BMS just commands the charge current to reduce based on cell group terminal voltage during charging, to ensure that no cell group exceeds the upper threshold voltage (max is 4.2 V per cell group, but it looks as if Tesla may use a more conservative value of around 4.15 V per cell group, perhaps (these voltages being for the non LFP cells, LFP is more than a little bit different).
It seems probable that the BMS not only logs the number of DC charges since the last cell group balancing event, but also logs the rate at which cell groups are going out of balance. If the pack is staying in balance, without the need for active balancing at the latter stages of an AC charge, then there would be no reason to throttle back rapid DC charging. On the other hand, if there is some voltage drift between cell groups, it seems probable that the BMS may well start to throttle rapid DC charging in order to protect the pack, and, perhaps, notify the owner that an AC charge is needed in order for the BMS to complete all its housekeeping work.
There is some speculation here, but most of it is based on my experience of building battery packs and battery management systems, and trying to determine what works well and what doesn't. In general, cell group balancing is much less of an issue with the topology that Telsa have adopted, with many cells connected in parallel within each cell group. Other topologies, like those with just one big cell in each cell group, seem very much more prone to going out of balance after only a handful of non-balancing charges.
Finally, rapid charging isn't rapid at all as far as cell C rate goes. Some power tools charge at higher C rates than even a V3 charger, as that only charges at about 3C, which is pretty feeble, and way less than the max discharge C rate. I've charged packs at 20 C before now, using older cell chemistry, too.
