Background: The Architect HIV Ag/Ab Combo has been implemented worldwide as a screening test for more than ten years. However, its high sensitivity may lead to false-positive results in a low-prevalence setting. Methods: This study was carried out to evaluate the performance of the Architect HIV Ag/Ab Combo based on the sample-to-cutoff (S/CO) ratios, to explore the optimal cutoff value to predict HIV infection and reduce the frequency of false-positive results. A retrospective analysis of clinical samples using the Architect HIV Ag/Ab Combo between July 2011 and February 2020 was performed. The sensitivity, specificity, positive predictive value (PPV), and false-positive rate (FPR) were evaluated and the receiver operating characteristic curve (ROC) analysis was used to determine the optimal cutoff value to predict HIV infection. Results: During the study period, 531 out of 692,155 samples were repeatedly reactive by the Architect HIV Ag/Ab Combo. The median S/CO value of HIV-positive were significantly higher than that of false-positive. The PPV of males (85.68%) was significantly higher than that of females (47.89%). The optimal cutoff value estimated by ROC analysis was 8.96 with the highest sum of sensitivity (100.00%) and specificity (100.00%) for males. However, for females, the optimal cutoff value was 26.97 with the highest sum of sensitivity (100.00%) and specificity (100.00%). Conclusions: For the Architect HIV Ag/Ab Combo, the optimal cutoff value needs to be set for the different genders to predict the final status of HIV infection reliably.