Isn't this the standard bimodal distribution we expect? Boys make up the majority of CEOs but they also make up the vast majority of criminals. Median girls are smarter than median boys, but the distribution is less wide on a population level.

That's an interesting statistic. Do you have a source? I'd be interested in learning more

The first is from a table of percentiles published by the College Board, I remembered that statistic off-hand, I think you find it by googling, or on their website. The second is something you can find somewhere, I know I've seen the raw data at some point. Some googling will at least find articles referring to that ratio, like this one: https://economics.mit.edu/files/7598

Interestingly enough, that has a graph of the AMC math contest breakdown by sex. Check out page 7 (page number 115). When you start getting in the upper percentiles there I think that culture might magnify the male/female ratio a bit (this is based on personal experience), but at least in my high school the center mass of the scores up to 100+ or so was filled out by teachers just signing up a bunch of good students with no particular involvement in mathematical extracurriculars.