The deep-sea Cenozoic planktonic microfossil record has the unique characteristics of continuously well-preserved populations of most species, with virtually unlimited sample size, and therefore constitutes, in principle, a major resource for macroevolutionary research. Antarctic Neogene radiolarians in particular, are diverse, abundant and consistently well-preserved and evolved rapidly. This fauna is, in theory, a near-perfect testing ground for paleodiversity reconstructions. In this study we determined the diversity history of these faunas from a new quantitative, taxonomically complete data set from Neogene and Quaternary sections at several Antarctic sites. The pattern retrieved by our whole-fauna data set shows a significant, largely extinctionless ecological change in faunal composition and decrease in the evenness of species' abundances during the late Miocene, followed 3 Myr later, at around 5 Ma, by a significant drop in diversity. We tentatively associate this ecological event with a synchronous, regional change in the composition of the primary producers, but as yet cannot identify any environmental changes associated with the later extinction. Further, our whole-fauna diversity history was compared to diversity computed from much less complete, biostratigraphically oriented studies of species' occurrences, compiled in the Neptune database and reconstructed by using subsampling methodologies. Comparison of our whole-fauna and subsampling-reconstructed diversity patterns shows that the first-order trends are the same in both, suggesting that, to some degree, such literature compilations can be used to explore diversity history of plankton. However, our results also highlight substantial errors and distortions in the reconstructed diversity which make it poorly suited to more-detailed studies (e.g., for comparison of diversity history with paleoenvironmental history). We conclude that detailed studies of plankton diversity, and particularly those attempting to understand the relation between diversity and paleoceanographic change, should be based on taxonomically comprehensive, quantitative data.