This paper evaluates the influence of the classmates' age on individual academic achievement. The identification strategy explores a mechanism for the division of classes based on age homogeneity, with cutoff value determined by the Brazilian law of age at school entry. Fuzzy regression discontinuity estimates a local average treatment effect of 2.34 standard deviations favorable to students assigned to class with older peers, in comparison to those allocated to classes with younger classmates. The empirical estimations use a unique educational dataset originated from the Brazilian Ministry of Education, which provides a large information set related to the student's scholar environment.