I think Reveillark is significantly more powerful.
What makes Reveillark so great is that it constantly brings threats into the table. It's a 4/3 flyer on itself but, if you kill it, it brings two more threats to the table. Worse of all, it may bring back Eternal Witness, which will return the Reveillark to play the next turn. It's a highly recursive creature and hard to deal with.
Compare what happens if you play either creature and your opponent wraths the board. With Reveillark, he's in as much of a pickle as before. With Sigardian Savoiur he is probably ahead.
I have a strong preference for
Reveillark for the reasons you discussed and more, but I don't think it's
significantly more powerful. This is a minor point but hear me out.
Assuming reasonable parity in the graveyard, the scenarios where Reveillark is absolutely better are where 1) either player casts a wrath, 2) your opponent uses non-exiling removal to not die from your flying elemental, 3) your opponent has a flying blocker, or, most importantly, 4) you have creatures in your yard with misleadingly low power in the graveyard that let you "cheat", like a
Mulldrifter,
Siege-Gang Commander,
Kalonian Hydra, or my favorite combo,
Karmic Guide.
To me,
Sigardian Savior is the better card when you're 1) behind on board, 2) you need to push through damage next turn to clinch the game, 3) your opponent has a Path/Swords/Settle/etc. to deal with your flying threat.
These scenarios where the Savior is better are less frequent but pretty important (especially stabilizing the board ASAP), but we're also talking about a hypothetical situation with a fully stocked graveyard, and this is where my biggest point of consideration comes up. Because the typical cube has many more 2-drops than it has 2-power cards, getting full value from
Sigardian Savior is simply going to happen more often. To this point, in my 720 list, there are
364 creatures with CMC 2 or less compared to just
201 creatures with power 2 or less.
The cards that technically meet one of those requirements requirements but will fizzle will fizzle in either case, such as with Hangarback Walker, so they've been left in those counts for convenience's sake.
For Reveillark, in
77 out of those 201, you get to "cheat" on mana relatively speaking (about 1/3 of the time), which is excellent for sure, but the scenarios where you're going to eke out the full value of bringing back two creatures with your white 5-drop is going to happen much more with the Savior when they make up 80% more of a cube list. This may not be true for your cube and is somewhat self-corrected by drafting/deckbuilding accordingly, but it's a pretty staggering difference in the possible pieces you can use imo, and more and more of white's premiere 2-drops are receiving a third point of power, as seen with two hotly discussed MID spoilers.
From this, I would argue Reveillark has a higher ceiling but a lower floor, and
maybe even a lower average case scenario. It's without a doubt more interesting, and the larger point is somewhat moot because I think cubes that run one of these cards will benefit from running both, given enough space, but I enjoyed this thought exercise and hope you did too.
Personally, I'm most happy about
Sigardian Savior because it gives me the best possible excuse to cut pet card holdout
Return to the Ranks, which is only run in 443 cubes on CubeCobra for good reason; even I have never successfully played it in my own cube.