There is no consensus among academics about whether children benefit from smaller classes. We analysed the data from the 2012 Programme for International Student Assessment (PISA) to test if smaller classes lead to higher performance. Advantages of using this data set are not only its size (478,120 15-year old students in 63 nations) and representativeness but also that the 2012 PISA data set, for the first time, includes the class size for each participating child. We found that, in most countries, children in smaller classes had a lower performance score in solving reading comprehension problems than those in larger classes. We further analysed the relationship between class size and factors that can explain this paradoxical phenomenon. Although grouping of students by ability and the socioeconomic status of parents played some role in some countries, these factors cannot fully explain the effect. We finish by discussing the overlooked potential advantages of larger classes.