High-throughput detection of neutralizing antibodies against SARS-CoV-2 presents a valuable tool for vaccine trials or investigations of population immunity. We evaluate the performance of the first commercial surrogate virus neutralization test (sVNT, GenScript Biotech) against SARS-CoV-2 plaque reduction neutralization test (PRNT) in convalescent and vaccinated individuals. We compare it to five other ELISAs, two of which are designed to detect neutralizing antibodies. In 491 pre-vaccination serum samples, sVNT missed 23.6% of PRNT-positive samples when using the manufacturer-recommended cutoff of 30% binding inhibition. Introducing an equivocal area between 15 and 35% maximized sensitivity and specificity against PRNT to 72.8-93.1% and 73.5-97.6%, respectively. The overall diagnostic performance of the other ELISAs for neutralizing antibodies was below that of sVNT. Vaccinated individuals exhibited higher antibody titers by PRNT (median 119.8, IQR 56.7-160) and binding inhibition by sVNT (median 95.7, IQR 88.1-96.8) than convalescent patients (median 49.1, IQR 20-62; median 52.9, IQR 31.2-76.2). GenScript sVNT is suitable to screen for SARS-CoV-2-neutralizing antibodies; however, to obtain accurate results, confirmatory testing by PRNT in a equivocal area is required. This equivocal area may require adaptation for use in vaccinated individuals, due to higher antibody titers.