This is further underlined when looking at the Swedish version of the SAT (‘Högskoleprovet’), which is used as an alternative way to get into university programs if your grades are not sufficient: in this test, men have consistently outperformed women. To eliminate the difference, changes were made to the test that would favor women, but eventually, the universities started complaining that the test no longer predicted academic performance well enough. When corrected, men again came out on top.

A possible explanation for the male advantage in the test is that more talented women are able to get into university on their grades alone, leaving the SAT slightly stacked with men whose test performance exceeded their grades.

BTW, subjective grading differences don’t seem to be a solely Swedish phenomenon.
Non-cognitive Skills and the Gender Disparities in Test Scores and Teacher Assessments: Evidence from Primary School (Cornwell et al 2011)

Boys score at least as well on math and science tests as girls, with the strongest evidence of a gender gap appearing among whites. However, boys in all racial categories across all subject areas are not represented in grade distributions where their test scores would predict. Even those boys who perform equally as well as girls on reading, math and science tests are nevertheless graded less favorably by their teachers, but this less favorable treatment essentially vanishes when non-cognitive skills are taken into account.

Even if one wants to argue that it’s appropriate to grade girls (and boys who are as well-behaved as the girls) higher, the fact remains that there seems to be a built-in difference in GPA that does not depend on cognitive ability – at least in this study looking at fifth-graders. In the Swedish report linked above, the trend grew stronger from 6th grade to 9th.