Here are the results of the Visual Intelligence test on SimilarMinds. (sample size 210k)

Male scores:

Female scores:

Males greatly outperform females on this test. Of test takers who obtained a perfect score, men outnumber women by a factor of three.

Here are the results of the Verbal Intelligence test on SimilarMinds.  (sample size 150k)

Male Results:

Female Results:

The test was taken by 63% women and 37% men, that's why all the bars are longer on the female results.  If you go by gender percentage,  a larger percentage of men obtained high scores.

If you look at the higher scores. You see the difference between male and female test takers gets larger at the highest test scores with almost twice as many men by percentage scoring 35 as women.

It's unclear what the explanation for these results are.     Since more women than men take internet tests in general, the smaller number of men that take online tests might be self selectively more intelligent than average.    A fairly similar argument has been made to explain the male/female performance gap on the SAT which is taken by more women than men.  However, according to this analysis, that theory does not hold up (at least for the math SAT).

A study which looked at 41 countries found males greatly outnumber females in high math scores and low reading scores whereas females greatly outnumber males in high reading scores and low math scores (source).      Reading tests measure comprehension though, the above test measures vocabulary.    I found another study which replicates the female comprehension advantage but shows a small male advantage for vocabulary (source).   I think I'll construct a reading comprehension test in the future and see what those results look like.

In any case, the pattern in the research I've read shows the female intelligence variance is smaller and performance distribution more bell shaped, male performance variance is larger with performance distribution more bowl shaped, so, more men with very low and very high IQs, more women in the middle. 
source

If this is true, what's the explanation?

According to some current genetics researchers, it is believed that intelligence is primarily inherited on the X chromosome (source).     Having two X chromosomes, women get an averaging out of two different intelligence profiles.  So if one is high, and one is low, they end up somewhere in the middle.  Men however just have the influence of one X.  If it's low, having no second X to average that out, they end up with low intelligence.   The authors of this idea link the genes associated with non specific mental retardation with intelligence.   There is evidence for non specific mental retardation being linked to the X chromosome, i.e. passed down by the mother (source 1, source 2).

Concluding, I would imagine that low intelligence males (and females) would be less likely to take intelligence tests on the internet.   Assuming there are more males with lower intelligence, my test results are not going to be able to pick up on that phenomenon.  However,  my results should be able to show any male or female cognitive advantages, if any exist (provided my male sample is not self selectively more intelligent than average, which they might be).  Asking for GPA (or other info) of test takers may allow me to determine how representative the male and female test takers are thereby enabling me to interpret the results in a more accurate context.