This paper presents the results of a study to compare the performance of design-bid-build (DBB), construction manager at risk (CMR), and design-build (DB) project delivery systems in the US building construction industry. By leveraging verified data from 212 projects, a best subset analysis was used to generate regression models that explain the greatest amount of variance in five measures of project performance: unit cost, cost growth, schedule growth, construction speed, and delivery speed. From these regression models, the average of the predicted project performance was calculated for each delivery system when holding all other variables constant. The results show that, on average, projects with a DB delivery system are delivered faster and with lower cost and schedule growth than projects structured as CMR or DBB. In addition, the completed unit cost of DB projects is comparable to DBB and slightly less than CMR projects. These results are generally consistent with findings from seminal articles published in the late 1990s, as well as more recent works that also compared performance across delivery systems. However, the modeling does indicate that, except for delivery speed, the gap in performance between DBB, CMR, and DB has narrowed over time. This research provides an updated benchmark for the performance of the most commonly used project delivery systems, which has value to owners seeking guidance when making strategic decisions in structuring their projects.