This paper proposes to shed new light on the mutations of the comprehensive schooling model by renewing the analysis of the effects of comprehensive education on student achievement. It does this by comparing across a wide range of outcomes between comprehensive and tracked systems. To accomplish this, the paper uses a newly developed dataset on the structures of lower secondary educational systems, student heterogeneity management tools, and the recent mutations of comprehensive systems. The dataset, comprising more than 20 quantitative and typological variables that cover the whole of the OECD countries, was constructed from existing international datasets (OECD, IEA, PISA, etc) and from national qualitative case studies when international data were not available. Using this dataset, analysis shows that the last twenty years were marked by the emergence of multiple models of comprehensive schooling which have permanently weakened the pertinence of the traditional binary typology opposing comprehensive and selective school systems. Further, the quantitative analysis developed within this paper shows that the performances of the selective school model are poor compared to comprehensive school models. It also shows that the three models of comprehensive schooling are associated with academic results that are different in terms of both effectiveness and equality.