Individual differences in academic success were investigated in a geographically defined whole-population sample of very preterm children with a gestational age of less than 32 weeks or a birth weight of less than 1500 gm. The sample consisted of 264 very preterm children (75.6% of German-speaking survivors) and 264 controls matched for gender, socioeconomic status, marital status and age of mother, who were studied from birth. The present analyses focused on the impact of cognitive skills assessed at ages 6 and 8 on academic success at the age 13. IQ scores, prereading skills, reading, spelling, and math performance assessed during the last kindergarten year and again at the end of Grade 2 were used as predictors of academic success in early adolescence. Differences between very preterm children and controls in cognitive abilities already observed in earlier assessments remained stable over time, with controls on average performing more than half a standard deviation above the level of preterm children. Preterm children also performed poorer on the literacy measures and indicators of math performance. Multivariate and causal modeling revealed different prediction patterns for the two groups. Whereas IQ was particularly important for the prediction of academic success in the pre-term sample, general IQ was less relevant for the prediction of academic success in the control group. When subgroups of at-risk children were formed according to birth weight categories, we found that school problems were most pronounced for children with extremely low birth weight (1000 gm and less). (Contains 4 figures and 4 tables.)