Empirical studies of high school mathematics typically report small gender differences in favor of boys. The present article challenges this established finding by comparing two competing structural conceptions of mathematical ability. The standard model assumes mathematical ability alone to account for the interindividual differences observed on the corresponding measures. The nested-factor model assumes interindividual differences in mathematical ability to be explained by two mutually independent factors, namely general cognitive ability and specific mathematical ability. The two models were examined using data from 29,171 ninth-graders. Results from the standard model replicated the typical finding of small gender differences in mathematical ability in favor of boys. However, results from the nested-factor model indicated large to very large gender differences in specific mathematical ability in favor of boys. Implications for research on gender differences in cognitive abilities are discussed.