HIV RNA plasma viral load (VL) is the standard surrogate marker to monitor response to antiretroviral treatment (ART). We compared the linearity, repeatability, and concordance of six commercially available HIV RNA VL platforms using clinical samples from patients from Brazilian sites where different HIV-1 subtypes co-circulate. A total of 150 plasma samples from each city were collected in Curitiba, Southern Brazil (subtype C), São Paulo (subtype B), and Santos (BF recombinants), Southeast Brazil. Platforms were VERSANT® Siemens HIV RNA 1.0 (kPCR); VERSANT® Siemens HIV-1 RNA 3.0 (bDNA); Abbott Real-Time HIV-1; NucliSens EasyQ® HIV-1 v2.0 Biomerieux; COBAS® TaqMan®, Roche; and artus HIV Virus-1 RT-PCR, QIAGEN. OptiQuant HIV-1 RNA quantification panel was used to compare VL linearity, using samples containing 50, 500,5,000, 50,000, 500,000, and 5,000,000 HIV copies/mL. HIV RNA panels with subtypes A, B, C, D, F, G, H, circulating recombinant form (CRF)1, and CRF2 were utilized. A high degree of linearity and repeatability was demonstrated for all platforms. When compared with a subtype B reference sample, 17 of 54 (31.48%) samples diverged by more than 0.5 log10 copies/mL. Except for the Roche platform, all platforms underestimated subtype C VLs. A total of 743 (82.6%) valid results were obtained with samples from São Paulo, 707 (78.6%) from Santos, and 673 (74.8%) from Curitiba (São Paulo vs. Santos, p = .03; São Paulo vs. Curitiba, p = .00006; Santos vs. Curitiba, p = .06). The number of discordant samples between different methodologies when VL was undetectable in one method and detectable in the other ranged from 1.25% (Abbot vs. Siemens) to 44.8% (Abbott vs. Biomerieux). Finding samples with undetectable VL in one method and a high VL in another might have important individual and public health consequences. Standardization of VL measurements, particularly for non-B subtypes infections, especially subtype C, is necessary to maximize the individual and public health benefits of ART globally.