The development and verification of HIV-2 assays depends in part on the availability of well-characterized samples, including those from reagent repositories. During the development of an HIV-2 RNA quantification assay, two HIV-2 viral isolates (CDC 301340 and CDC 301342) obtained from the NIAID AIDS Reagent and Reference Repository were not detected leading to an investigation. Two HIV-2 primers/probe sets of known performance in real-time viral RNA quantification assays, targeting different regions of the virus, also failed to generate RT-PCR products for these two isolates. These isolates were tested in the HIV-1 specific COBAS AmpliPrep/COBAS TaqMan HIV-1 Test v2.0 (Roche Molecular Diagnostics) and were quantified at high copy number. Other HIV-2 isolates tested were not amplified in the COBAS HIV-1 TaqMan assay. Furthermore, the discrepant isolates were highly reactive in an HIV-1 p24 antigen test while the other HIV-2 isolates showed very weak, if any, cross-reactivity with the HIV-1 p24 assay. Phylogenetic tree analysis of sequences from the protease-reverse transcriptase regions of the discrepant HIV-2 isolates mapped with HIV-1 Group M, Subtype CRF02_AG confirming these isolates were of HIV-1 origin and had been misclassified as HIV-2. The use of misclassified isolates in the verification of molecular and immunological assays can lead to misinterpretation of test results, misdirection of efforts into assay redesign and increased development costs. The results of this study were shared with the NIAID AIDS Reagent Program, leading to the reclassification of the two discrepant isolates as HIV-1.