What's this about A5s?

It makes sense to have some bluffs to balance our value hands, and we want to block the top of our opponent's range when we're bluffing.

If you accept those two premises, then you need to control which combos you use to bluff with so that you don't overdo it. A5s and A3s aren't drastically different, but if you bluff with every small suited ace in narrow range spots you'll be overbluffing in comparison to how many value hands you have.

So you choose the best possible combos to use as your first bluffs. A5s is slightly better so it makes sense to choose that combo first. Every straight has either a T or a 5 in it, which gives A5s anywhere from an extra tenth of a percent up to a full 1% or so of additional equity compared to other weak suited aces when you're up against ultra-tight ranges. So mathematically it makes sense to bluff with A5s before A3s.

That's not to say A3s is a bad bluff candidate either. In practice it might make more sense to use live reads or something else to differentiate which combos to bluff with, rather than choosing based on our specific combo, but you don't want to overdo it with your bluffs so you need some kind of system to control your combos, assuming your opponents will punish you if you get out of line and bluff too often.

Great White states it well, it’s all about frequencies. You could pick A3 and use it instead of A5, but you don’t want to bluff both those hands or you’ll be bluffing too much.

The solver gives you guidelines, but each situation is different. A lot of times at low stakes, you don’t want to be bluffing at all.

The A5s vs A3s distinction makes even more sense when you factor in specific board textures. On boards like 2-3-4 or 3-4-6, A5 picks up gutshot equity that A3 doesn't, which is part of why solvers prefer it as a bluff. It's not just blocking, it has more ways to improve if called.