Re: One drag model more accurate then another?
Mathematically and theoretically - for Sierra MK, Lapua Scenar and Berger VLD (and similar) bullets G7 drag model is the appropriate <span style="text-decoration: underline">standard</span> model (because they are similar to the standard Type 7 Projectile, and very dissimilar from the standard Type 1 Projectile).
As for the actual implementations - since you probably won't be writing your ballistic calculator yourself - <span style="text-decoration: underline">regardless of the drag model or BC type used</span> some programs tend to be more accurate than others. Programs such as LB3 or QTU build their own (non-standard) drag curves based on some wisdom of their authors.
Thus - if you want to compare projectiles based on their BC, then G7 for the modern rifle bullets is the right one to use. <span style="font-style: italic">Though majority of the bullets on the market come only with G1 BC, and it's quite inflated & hardly usable for <span style="text-decoration: underline">long range</span> predictions. Bryan Litz in his book book gives (banded!) G7 BC for several long range bullets - but there are far more bullets than Bryan could test, measure and document.</span>
If you just want to compute trajectories for different bullets (and maybe compare them this way) - then what drag model the program uses doesn't matter <span style="text-decoration: underline">because the authors likely tweaked those models almost beyond recognition</span> (the Army term is FUBAR - which stands for Forced Unrecognizable By Author's Redefinition [of the standard]
). <span style="font-style: italic">And if they didn't tweak them that far - there'd be no way in hell such a program could accurately predict bullet drop for e.g. 308Win 175gr SMK at 1000m using G1 BC. Which of course matters little to a shooter who just needs his prediction data and couldn't care less how it is computed as long as its accurate = matches reality.</span>