Most of the buildings in Turkey constructed before 1999 Kocaeli and Duzce Earthquakes do not have sufficient seismic resistance due to construction errors and do not satisfy the recent Turkish Building Earthquake Code (2018) or its preceding versions. In Istanbul, the building stock consists of 1,116,300 buildings. Clearly, seismic safety assessment of this amount of buildings is not practically possible by code-based detailed assessment approaches, which generally assess the seismic performance of buildings in terms of pass or fail, rather than rank and classify/prioritize these buildings considering their seismic safety. Moreover, the probability of a catastrophic earthquake that could hit Istanbul is high. Therefore, to minimize casualties and economic losses, it is important to determine the most vulnerable existing buildings effectively in terms of cost and duration. Accordingly, Istanbul Metropolitan Municipality has started to evaluate the seismic safety of existing reinforced concrete buildings using the PERA2019 method considering the actual structural data collected on site. The buildings are then divided into 5 seismic risk classes from low-risk to very high-risk according to the computed seismic safety ratio. In this study, the reliability of the PERA2019 is evaluated by considering the predicted risk classes of another large group of actual buildings that have been subjected to actual earthquakes and experienced earthquake-induced damages in different extents. The comparative evaluation of actual seismic damages and predictions of PERA2019 showed clearly that the methodology has a sound ability to estimate the structural seismic performance of the investigated buildings.