Purpose: The purpose of this study was to measure the variability of torque produced by a population of mechanical torque-limiting devices in clinical service in a US dental school. The torque-limiting devices were divided into two categories according to their mode of action: toggle-type and beam wrenches. Proper action of these devices is essential for calibrated delivery of preload to implant prosthetic screws. Materials and Methods: Seventeen torque-limiting devices (35 Ncm) were obtained from graduate prosthodontic, predoctoral, and faculty practice clinics. Nine of these were toggle-type devices, and eight were beam-type wrenches. Torque from each wrench was measured using an MGT electronic torque meter. Wrenches were tested in two modes, slow (over 4 seconds) and fast (over 1 second). Results: Toggle-type torque wrenches produced a mean (± SD) torque of 38.1 ± 16.0 Ncm; beam-type wrenches produced 32.8 ± 1.1 Ncm. These results were not significantly different. When tested in fast mode (1 second), toggle-type wrenches produced 28.0 ± 9.6 Ncm; in the slow mode (4 seconds) they produced significantly more force, 36.6 ± 14.0 Ncm (p < 0.001). Beam-type wrenches produced 33.2 ± 1.1 Ncm and 32.8 ± 1.1 Ncm in fast and slow modes, respectively. Conclusions: Both types of wrenches tested were capable of producing accurate torque values; however, variability was higher in the toggle-type group. Some toggle-type torque wrenches in clinical service delivered unacceptably high torque values. It is recommended that clinicians calibrate toggle-type wrenches frequently. Torque wrenches should be activated slowly, over 4 seconds, when using a correctly calibrated toggle-type wrench. © 2009 by The American College of Prosthodontists.