Background Antigen rapid diagnostic tests (RDT) for SARS-CoV-2 are fast, broadly available, and inexpensive. Despite this, reliable clinical performance data is sparse. Methods In a prospective performance evaluation study, RDT from three manufacturers (NADAL, Panbio, MEDsan) were compared to quantitative reverse transcription polymerase chain reaction (RT-qPCR) in 5 068 oropharyngeal swabs for detection of SARS-CoV-2 in a hospital setting. Viral load was derived from standardized RT-qPCR Cycle threshold (Ct) values. The data collection period ranged from November 12, 2020 to February 28, 2021. Findings Overall, sensitivity of RDT compared to RT-qPCR was 42.57% (95% CI 33.38%-52.31%), and specificity 99.68% (95% CI 99.48%-99.80%). Sensitivity declined with decreasing viral load from 100% in samples with a deduced viral load of 10^8 SARS-CoV-2 RNA copies per ml to 8.82% in samples with a viral load lower than 104 SARS-CoV-2 RNA copies per ml. No significant differences in sensitivity or specificity could be observed between the three manufacturers, or between samples with and without spike protein variant B.1.1.7. The NPV in the study cohort was 98.84%; the PPV in persons with typical COVID-19 symptoms was 97.37%, and 28.57% in persons without or with atypical symptoms. Interpretation RDT are a reliable method to diagnose SARS-CoV-2 infection in persons with high viral load. RDT are a valuable addition to RT-qPCR testing, as they reliably detect infectious persons with high viral loads before RT-qPCR results are available. Funding German Federal Ministry for Education and Science (BMBF), Free State of Bavaria