The G7 profile does not necessarily have "less drag" as far as the ballistic calculator is concerned. That totally depends on the relative G1 and G7 ballistic coefficient values that are actually input into the program. In his "Ballistic Performance of Rifle Bullets, 3rd Ed.", Litz gives respective values for the
average G1 and G7 BCs or 0.457 and 0.234, respectively, between 3000 and 1500 fps. The values you used (0.530 per Hornady and 0.246 avg. per Litz) were determined by independent groups and are much farther apart, indicating the
relationship between the two different BC values may not be correct.
If you compare G1 and G7 BCs determined from consistent velocity data sets as Litz has done in his book, what you will find is that multiplying the G1 BC for almost any bullet listed in the book by a value very close to 0.51 will generate the G7 value. I believe this relationship is related to the ratio of the form factors of the G1 and G7 standards. Regardless, the ratio of the two values you used is much lower, suggesating one of the two BC values was off relative two the other (0.246/0.530 = ~0.46, instead of ~0.51). I would guess that the Hornady value of 0.530 is inflated relative to its true value, which therefore "tricks" the ballistic calculator into thinking it has less drag than it really does, and therefore greater retained velocity than the same calculation made with the G7 BC value.
Nonetheless, if Litz' measured G1 and G7 Ballistic coefficients of 0.457 and 0.234 are used at JBM Ballistics, the program still predicts the bullet to have greater retained velocity at 1000 yd with the G1 coefficient. The values may be a bit closer than when using Hornday's G1 BC, but they're still not exactly the same, differing by ~42 fps.
However, if you run a similar comparison using Litz' G1 and G7 values for the Berger 90 VLD, the outputs are very close, as would be expected. I did the same thing with a .30 cal bullet, and the results were also very close, as they should be.
I cannot explain the discrepancy between G1 and G7 outputs with the Hornady 105 bullet. However, one would think it almost has to be rooted in a greater difference between the two G1/G7 BC values measured by Litz than there actually is. In other words, one of the two values must somehow be off. Otherwise, it would be expected that velocity decreases predicted over distance using the two different BCs would be very close. Not identical, perhaps, but very close.