Human immunodeficiency virus types 1 and 2 (HIV-1 and HIV-2) diagnostic testing algorithms recommended by the Centers for Disease Control involve up to three tests and rely mostly on detection of viral antigen and host antibody responses. HIV-1 p24 antigen/HIV-1/HIV-2 antibody-reactive specimens are confirmed with an immunochromatographic HIV-1/HIV-2 antibody differentiation assay, and negative or indeterminate results from the differentiation assay are resolved by an HIV-1-specific nucleic acid amplification test (NAT). The performance of a proposed alternative algorithm using the cobas HIV-1/HIV-2 qualitative NAT as the differentiation assay was evaluated in subjects known to be infected with HIV-1 (n = 876) or HIV-2 (n = 139), at low (n = 6,017) or high (n = 1,020) risk of HIV-1 infection, or at high-risk for HIV-2 infection (n = 498) (study A). The performance of the cobas HIV-1/HIV-2 qualitative test was also evaluated by comparison to an HIV-1 or HIV-2 alternative NAT (study B). The HIV-1 and HIV-2 overall percent agreements (OPA) in study A ranged from 95% to 100% in all groups. The positive percent agreements (PPA) for HIV-1 and HIV-2 were 100% (876/876) and 99.4% (167/168), respectively, for known positive groups. The negative percent agreement in the HIV low-risk group was 100% for both HIV-1 and HIV-2. In study B, the HIV-1 and HIV-2 OPA ranged from 99% to 100% in all groups evaluated (n = 183 to 1,030), and the PPA for HIV-1 and HIV-2 were 100% and 99.5%, respectively, for known positive groups. The cobas HIV-1/HIV-2 qualitative assay can discriminate between HIV-1 and HIV-2 based on HIV RNA and can be included in an alternative diagnostic algorithm for HIV.