OBJECTIVES: We describe neuropsychological test performance (NP) in antiretroviral treatment (ART)-naive HIV-positive individuals with CD4 cell counts above 500 cells/muL. METHODS: In a neurology substudy of the International Network for Strategic Initiatives in Global HIV Trials (INSIGHT) Strategic Timing of AntiRetroviral Treatment (START) study, eight neurocognitive tests were administered. The primary measure of NP was the quantitative NP z-score (QNPZ-8), the average of the z-scores for the eight tests. Associations of baseline factors with QNPZ-8 scores were assessed by multiple regression. Mild neurocognitive impairment (NCI) was defined as z-scores < -1 in at least two of six cognitive domains. RESULTS: A total of 608 participants had a median age of 34 years; 11% were women and 15% were black; the median time since HIV diagnosis was 0.9 years; the median CD4 cell count was 633 cells/muL; 19.9% had mild NCI. Better NP was independently associated with younger age, being white, higher body mass index (0.10 per 10 kg/m2 higher), and higher haematocrit percentage (0.19 per 10% higher). Worse NP was associated with longer time since HIV diagnosis (-0.17 per 10 years), diabetes (-0.29) and higher Framingham risk score (-0.15 per 10 points higher). QNPZ-8 scores differed significantly between geographical locations, with the lowest scores in Brazil and Argentina/Chile. CONCLUSIONS: This is the largest study of NP in ART-naive HIV-positive adults with CD4 counts > 500 cells/muL. Demographic factors and diabetes were most strongly associated with NP. Unmeasured educational/sociocultural factors may explain geographical differences. Poorer NP was independently associated with longer time since HIV diagnosis, suggesting that untreated HIV infection might deleteriously affect NP, but the effect was small.