I have never seen any authoritative information regarding a requirement for balancing Tesla batteries. Random Internet strangers are not authoritative.
It is very clear from the car behaviour that balancing occurs at 100% (or more specifically, during the variable but sometimes very long period when charging between 99% displayed on the gauge and the charge stopping at 100%). The car sits in this condition for long (but variable) periods of time consuming a very small charge current. It is also known from battery tear-down photos that the pack contains conventional bypass resistor balancing circuits.
We don't know from a completely authoritative source that it is definitely doing top balancing, but this is common industry practice and from evidence visible to any owner it looks very much like the car is doing it; no evidence or theory has been offered to contradict this. So I consider this true beyond reasonable doubt. Less certain is whether it attempts to do any balancing at lower charge levels. It doesn't appear to do general balancing during regular 90% charging (as otherwise you wouldn't have accumulated imbalance to give the variable time spent balancing on the rare occasion you then do a 100% charge); possibly it might correct gross imbalance during regular charging and only leave fine balance for a 100% charge.
Then there is the question of whether (or to what extent) imbalance actually matters.
It's very clear from the physics of the situation that imbalance gives you lower available capacity/range, but this is not a permanent loss as it is immediately restored just by doing the balancing. In the Tesla case, this really doesn't matter much: if you are charging to 90% then you plainly don't care about getting the ultimate range that day (else you would have charged to 100%), and on the day when you do want ultimate range you will have charged to 100% and so the battery will be balanced. So the only time it has any impact at all is if you have been charging to 90% for months and then suddenly decide that you want to make a testing-the-limits-of-range trip and you happen to have exactly 90 minutes of charging time before departure: that's enough to top off from 90% to 100% on a home chargepoint, but not enough to be sure of completing balancing. But that's an exceedingly unlikely scenario - much more likely your trip is planned in advance so you charge to 100% overnight, or you need to leave "right now" and don't have time to take advantage of the little bit of extra capacity that balancing would have given you. So essentially this can be ignored.
Separately, it is likely that imbalance will lead to inaccuracy in the 'fuel gauge'. From the physics, it's unlikely that this effect doesn't exist at all; from practical experience it seems like the effect is small.
Finally, there's the question of whether driving with a state of imbalance leads to more wear/permanent degradation than driving with the pack perfectly balanced. The theory that says this is important is that imbalance causes the worst-case cells to charge to more extreme voltages (high or low) than the ones in the middle and that more degradation occurs under those conditions; furthermore, that this has a cumulative effect because those worst case cells suffering the most wear are likely to become even more worst-case. This is by far the most speculative aspect of this discussion, as we don't have any authoritative information about what causes degradation, other than that Tesla tells us 100% should be avoided while 90% is OK. Plainly, if the theory is correct then the greater the extent of the imbalance the more it matters. Anecdotal results (looking for example at the amount of 'gauge range' recovered by doing a balancing cycle, or the time taken to balance against the teardown pics view of the balancing resistors which would take >20 hours to balance by 1% of capacity) suggest that imbalance is seldom more than 1%. So while the effect might be there, it's not huge.
My personal opinion from all of this is that doing a charge to 100% maybe once a year might be beneficial - but if you do at least one long trip a year you will have done that without having to think about it.
Finally, it's worth noting that there is no point in stressing over the absolute best charging strategy, as it is impossible to achieve it! Even if you knew everything about the cell chemistry and the behaviour of the car, you still can't do it unless you know in advance exactly how much you will be driving every day. One unplanned trip and your strategy is toast.