It actually depends on how strong the opponent's range is. Although interestingly not in a linear way. For example, if you push all-in vs...
- the bottom 50% of hands (<- this is never relevant but just for demonstration), it's 67.4% vs. 83.3% equity (always listing AKs) first
- a 100% range, it's 67.1% vs. 77.5%
- a top 25% range, it's 65.4% vs. 66.3%
- a top 15% range, it's 63.2% vs. 61.6%
- a top 10% range, it's vs. 59.1% vs. 58.1%
- a top 5% range, it's 48.9% vs. 52.5%
- a top 3% range, 42.4% vs. 45.1%
The start makes a lot of sense, e.g., AKs has only 69.1% against 27o whereas JJ has 88.2%, so it makes sense that JJ crushes the weak ranges much more. I don't know why AKs peaks somewhere between 10% and 15% though, I'd have expected that AKs just does relatively better the stronger the opponent's range is, mostly because it blocks AA and KK. Interesting...
But the ranges where AKs does best are also some of the most relevant. E.g., a 15% range looks like this:
This is a pretty typical shoving range. Though then again, I guess you're always shoving both hands here so it doesn't matter. The only cases where it would be close is if a range gets even stronger, and then JJ does better again. So I guess JJ is the winner; there are rare cases where you should shove JJ but fold AKs (like the top 5% range!), whereas it seems like there is essentially no situation where you should ever shove AKs but fold JJ.