The primary aim of this study was to assess the risks of fetal growth restriction and birth defects in children exposed prenatally to newer and older antiepileptic drugs, using an unselected epilepsy cohort. Deliveries recorded in the compulsory Medical Birth Registry of Norway 1999-2011 formed the study population. All 2,600 children exposed to antiepileptic drugs during pregnancy were compared to all 771,412 unexposed children born to women without epilepsy. Children of untreated mothers with epilepsy served as an internal control group. The main outcomes were small for gestational age birth weight and head circumference, and major congenital malformations. Children exposed to antiepileptic drugs had a moderate risk of growth restriction. Infants exposed to topiramate had a considerable risk of microcephaly (11.4 vs. 2.4 %; OR 4.8; CI 2.5-9.3) and small for gestational age birth weight (24.4 vs. 8.9 %; OR 3.1; 95 % CI 1.9-5.3). Carbamazepine, lamotrigine, levetiracetam, oxcarbazepine, gabapentin, and pregabalin had low malformation rates, whereas topiramate tended to have an elevated malformation rate. Valproate monotherapy was associated with a significant risk of birth defects (6.3 vs. 2.9 %; OR 2.5; CI 1.6-3.8), and specifically with septal heart defects and hypospadias. For mothers using valproate, the presence of major birth defect in one child was associated with a markedly increased risk for the siblings (42.9 vs. 6.7 %; OR 10.4; CI 2.3-46.7). Children of untreated mothers with epilepsy had malformation risk similar to the reference group. In conclusion, topiramate was associated with a substantial risk of fetal growth restriction, and possibly an increased malformation rate. Other newer-generation antiepileptic drugs had a low malformation rate. Valproate monotherapy had a significant malformation risk, especially in repeated pregnancies.