Reverse transcriptases (RTs) are typically assayed in vitro with 5-10 mM Mg2+, whereas the free Mg2+ concentration in cells is much lower. Artificially high Mg2+ concentrations used in vitro can misrepresent different properties of human immunodeficiency virus (HIV) RT, including fidelity, catalysis, pausing, and RNase H activity. Here, we analyzed nucleoside (NRTIs) and non-nucleoside RT inhibitors (NNRTIs) in primer extension assays at different concentrations of free Mg2+. At low concentrations of Mg2+, NRTIs and dideoxynucleotides (AZTTP, ddCTP, ddGTP, and 3TCTP) inhibited HIV-1 and HIV-2 RT synthesis less efficiently than they did with large amounts of Mg2+, whereas inhibition by the "translocation-defective RT inhibitor" EFdA (4'-ethynyl-2-fluoro-2'-deoxyadenosine) was unaffected by Mg2+ concentrations. Steady-state kinetic analyses revealed that the reduced level of inhibition at low Mg2+ concentrations resulted from a 3-9-fold (depending on the particular nucleotide and inhibitor) less efficient incorporation (based on kcat/Km) of these NRTIs under this condition compared to incorporation of natural dNTPs. In contrast, EFdATP was incorporated with an efficiency similar to that of its analogue dATP at low Mg2+ concentrations. Unlike NRTIs, NNRTIs (nevirapine, efavirenz, and rilviripine), were approximately 4-fold (based on IC50 values) more effective at low than at high Mg2+ concentrations. Drug-resistant HIV-1 RT mutants also displayed the Mg2+-dependent difference in susceptibility to NRTIs and NNRTIs. In summary, analyzing the efficiency of inhibitors under more physiologically relevant low-Mg2+ conditions yielded results dramatically different from those from measurements using commonly employed high-Mg2+ in vitro conditions. These results also emphasize differences in Mg2+ sensitivity between the translocation inhibitor EFdATP and other NRTIs.