The incremental validity of specific cognitive abilities beyond general intelligence has been investigated in studies using hierarchical multiple regression analyzes (HMR). In the present study, we investigated whether the incremental validity of specific cognitive abilities (i.e., verbal, figural, and numerical ability) for the explanation of school grades varied as a function of students' general ability level. Data comprised the standardization sample (N = 1371, ages 12.5-16.5) of the Berlin structure-of-intelligence test (BIS-HB; Jäger et al., 2006). Results of HMR analyses revealed that the specific cognitive abilities had little incremental validity in the low- and mid-IQ group. However, they provided significant benefit for the explanation of school grades in the high-IQ group. The results suggest that the interpretation of intelligence tests may benefit from a differential weighting of different test information, depending on general ability of the tested individual.