Introduction: Genes have been suggested to interact with predictors of school performance, but evidence is scarce. The purpose was to examine whether a hyperactive temperament leads to different school performance, depending on variability in a serotonin receptor gene (5-HTR2A). Method: The participants were a population-based sample of 909 girls and boys in Finland aged 3-12 when their mothers rated them on the temperament dimension activity (low in activity, average in activity, or restless/hyperactive). Genomic DNA was extracted from peripheral blood samples. There were three 5-HTR2A allelic groups: T/T, T/C, and C/C. School performance was measured by grade point averages (GPA) in elementary school at the age of 12. Results: The general linear model revealed a significant interaction between activity and the 5-HTR2A polymorphism (p=0.019). High activity predicted lower GPA in the T/T genotype (b=-0.25, p=0.003), and to some extent in the C/C genotype (b=-0.14, p=0.002) while no association between temperament and GPA was found in carriers of the T/C genotype (b=0.03, p=0.600). The T/T group with high activity has poorest GPAs compared to the other allelic groups. Conclusion: A temperament characteristic that is traditionally considered as risky may have different educational outcomes, depending on genetic vulnerability. The social or biological mechanisms whereby genes may moderate educational outcomes should be studied further. (Contains 2 tables and 1 figure.)