We have evaluated two commercially available kits (AMPLICOR MONITOR [Roche] and NASBA HIV-1 QT or NucliSens HIV-1 QT [Organon Teknika]) and two noncommercial methods for the accurate quantitation of human immunodeficiency virus type 1 (HIV-1) RNA in seminal plasma. The same panels of coded specimens were tested on four separate occasions. Laboratories using the commercial assays employed silica beads to isolate HIV-1 RNA, which removed inhibitory factors sometimes found in seminal plasma. Sensitivities and specificities, respectively, for each assay were as follows: AMPLICOR MONITOR, 100 and 73%; NASBA HIV-1 QT, 84 and 100%; NucliSens HIV-1 QT, 99 and 98%; and noncommercial assays, 91 and 73%. When results from the laboratory that was inexperienced with the silica bead extraction method were excluded from the analysis, specificity for the Roche assay increased to 100%. The commercial assays demonstrated highly reproducible results, with intra-assay standard deviations (measured in log(10) RNA copies/milliliter of seminal plasma) ranging from 0.11 to 0.32; those of the noncommercial assays ranged from 0.12 to 0.75. Differences in mean estimated HIV-1 RNA concentrations were </=0.67 log(10) and were greater at low viral loads. Suspension matrices that used blood plasma or seminal plasma did not make a difference in recovery of HIV-1 RNA, which suggested that blood plasma specimens can be used as external controls for seminal plasma assays. More variation in the HIV-1 RNA viral loads was observed in the seminal plasma values than in the blood plasma values when paired specimens from HIV-1-infected men were tested. Quantitation of HIV-1 RNA in seminal plasma can be reliably accomplished using two commercially available assays, and may be incorporated into the evaluations of HIV-1 seropositive men enrolled in clinical studies.