The sorting of high-ability workers is often advanced as one source of spatial disparities in economic outcomes. There are still few papers that analyze when human capital sorting occurs and whom it involves. Using data on 16 cohorts of university graduates in Sweden, we demonstrate significant sorting to urban regions on high school grades and education levels of parents, i.e., two attributes typically associated with latent abilities that are valued in the labor market. A large part of this sorting has already occurred in deciding where to study, because the top universities in Sweden are predominantly located in urban regions. The largest part of directed sorting on ability indicators occurs in the decision of where to study. Even after controlling for sorting prior to labor market entry, the “best and brightest” are still more likely to start working in urban regions. However, this effect appears to be driven by Sweden's main metropolitan region, Stockholm. We find no influence of our ability indicators on the probability of starting to work in urban regions after graduation when Stockholm is excluded. Studies of human capital sorting need to account for selection processes to and from universities, because neglecting mobility prior to labor market entry is likely to lead to an underestimation of the extent of the sorting to urban regions.