If I am literally guaranteed 5th in save percentage, then $13M+ is a no brainer because you can game it. That goaltender will maintain exactly the same save percentage regardless of shot quality, so you can instill a system that heavily limits shot quantity and pays no mind to shot quality.
Think of something like San Jose’s defense last season which finished 2nd from last in shots against but only 14th from last in expected goals against, and consider that that team was not even trying to game the system at all. If you actually tried, you could strategically do an incredible job of limiting shot quantity while giving up as much quality as you want because quality literally does not matter.
The goalie with the 5th best SV% last season was Thomas Greiss at 92.66%. That save percentage is 1.66% above average. Based on how goals saved above average is calculated, average SV% last year was 91%. A goaltender playing 70 games behind SJ’s defense would give you about 33 GSAA, but then you have to consider that one GSAA behind SJ’s defense is worth about 1.5 GSAx (this figure is almost exact if you look at GSAA/GSAx for their goalies) because of how high the quality of shots was that they allowed. So, right off the bat, a goaltender with Thomas Greiss’ SV% on last year’s SJ team, playing 70 games, is worth about 50 goals.
Now if you tailor your system to be a system that allows insane shot quality but low shot quantity, you can probably get significantly more than 50 goals saved above what a truly average goaltender would. Even at just 50 goals saved above an average goaltender, you’re still looking at by far the best player in the NHL if your goaltender can consistently maintain that.
If we look at things more practically and in line with what the OP is asking for and say we’re talking about the 5th best “goaltending performance”, and that this goaltender is only human like any other and their SV% can be influenced by team performance, then things get a lot more dicey. If they play 70 games per season, the 5th best per-game goaltending performance is probably saving you about 25-30 goals more than an average goaltender would based on metrics like GSAx. Based on metrics like Evolving Hockey’s GAA, only two skaters (Datsyuk and McDavid) have ever been worth over 25 goals above average in one season. While I think those metrics kind of under rate the value of truly elite players, I think it’s a pretty good baseline to work off of. And I think it’s reasonable to say that the absolute best player, in any given year, is probably not worth too much more than 25-30 goals above average, and that not more 2-3 players are worth that in any given season.
So, if you’re getting a goaltender whose performance is consistently posting the 5th best per-game goaltending performance, and they’re consistently playing 70 games, you’ve got a top-5 player in the NHL in just about every season. And by virtue of consistently being a top-5 player in every single season, that goaltender is probably the most valuable player in the NHL. Worst case, they’re 2nd behind Connor McDavid.
Regardless of how you interpret the question, I think that over $13M is the best answer.
