Objective. The aim of this study was to quantitatively assess the performance of metal artifact reduction (MAR) algorithms on the volume of metal cylinders, considering the influence of materials, positions, and fields of view (FOVs), by using 3 cone beam computed tomography (CBCT) devices (NewTom VGi evo, Picasso Trio, and ProMax 3-D Max). Study Design. Nine phantoms containing cylinders of amalgam, copperÀaluminum (CuAl) metal alloy, and titanium, combined in up to 3 positions, were scanned by using 2 different FOVs. MATLAB software was used to evaluate the differences between volumes before and after MAR application, and the possible interference of materials, positions, and FOVs. Wilcoxon's test and the Kruskal-Wallis test were used at a level of significance of 5%. Results. In general, images containing amalgam and CuAl showed a significant difference in volume before and after MAR application. However, no significant difference after MAR was observed (P > .05) relative to positions and FOVs. MAR had an impact on the cylinder volumes only in the NewTom VGi evo and ProMax 3-D Max scanners. Conclusions. The performance of MAR algorithms in volume correction of metal objects is dependent on the materials and the CBCT unit. (Oral Surg Oral Med Oral Pathol Oral Radiol 2020;000:1À8) Statement of Clinical Relevance This study confirmed the variability in MAR algorithm performance according to the tested CBCT devices and different types of materials, and the inaccurate performance of all of them compared to a gold standard measurement.