B > A, bigger range is harder / more commital (esp. with minions) to defend. Or described from other side, safer to poke / provoke reaction.

We already have demon2 / demon3, which can after 1 move use the unblockables to fork minion line, but not backline. This option / reason is not good enough for d2/3 to see play (it is played, just not for rushing initial forks). Having the backline forks after 1 move - not so often would enemy have his 3rd row undefended. So I think A is much less scary as rush. Also, the extra teleports / swaps on demon3 compensate for the extra fork after 1 move.

On other hand, when we had valkyrie before the action limitation, it was used a lot for rush, and angel is too. B is less tradeable ofc, but opponents 4th is defended much less often. Starting as "frozen" (3 turn inaction) helps to neutralise the rush, but even with that I would be afraid of A more from the start of match.

For me, attached to rook, the B unblockables are harder to predict and more dangerous than bishop (B vs queen0). Maybe not so much for better players, but I expect with lower skill, or higher ceiling to predict, they are even more so. Which shows that I don't consider myself near the top players, and therefore my view / feel of balance might differ from them. How much of this advantage is compensated by the initial "freeze" ? At around my rank, I don't think it does very much, I think the surprise factor can sneak in easier when the position gets more complicated, rather then straight king rush from start. Still it does help if the opponents setup is not a standard, lacking option to react adequately in 1 / few moves.

I consider demon3 <= A < queen0, A about 18 / 19.
I consider angel2, queen0 < B < angel 3, B about 22.