To compare ex vivo the accuracy of three electronic apex locators (EALs): Root ZX, Elements Diagnostic Unit and Apex Locator and ProPex. Electronic working length determination was carried out in 40 extracted teeth using an ex vivo model. After access preparation, a first operator determined the reference length (AL) for each tooth under a 30x stereomicroscope using the apical constriction as the apical landmark. All teeth were then measured with each EAL and the results obtained were compared with the corresponding AL. The AL was subtracted from the electronically determined distance. The measurements exceeding the AL were recorded as positive (long) and the measurements short of the AL were recorded as negative. Data were analyzed using the Friedman Test and Tukey multiple range test for nonparametric correlation amongst groups. Statistical significance was considered at P < 0.05. Comparing the differences between measurements obtained with the three EALs and those obtained with the stereomicroscope, the percentage of measurements within +/-0.5 mm of the AL was 97.37% (84.22% within 0.5 mm short of AL) for the Root ZX, 94.28% (88.57% within 0.5 mm short of AL) for the Elements and 100% (35.9% within 0.5 mm short of AL) for the ProPex. The mean difference between the AL and the lengths measured by the Root ZX, the Elements and the ProPex were, respectively, -0.157 +/- 0.228, -0.103 +/- 0.359 and 0.307 +/- 0.271 mm. The results of the present study confirm that the EALs determined the canal length within +/-0.5 mm from the apical constriction in the majority of cases. The majority of the ProPex readings were long.