Grouping students by academic achievement level has been practised in a wide variety of forms and contexts for over a century. Despite a general consensus in the research that between-class achievement grouping provides no overall benefit for students, the practice has persisted in various guises. Between-class achievement grouping is common in high schools, and is also practised in a number of primary schools in various countries. While the affective outcomes of such practices have been investigated recently, academic outcomes at primary level have not been studied in recent decades. This paper examines the academic outcomes of between-class achievement grouping in literacy and numeracy classes in Australian primary schools. Results from standardised tests are compared between two groups of schools--one regroups students for these areas, and one maintains mixed-achievement classes. It is argued that the current regrouping practice closely resembles streaming and provides no apparent academic advantage for students.