Objective: To measure the deviations of four different cone beam computed tomography (CBCT) devices in three dimensions by means of a three-dimensional (3D) implant-planning program. Materials and methods: A master radiographic template with two vertical, two transverse, and two sagittal radiopaque markers was fabricated for a human dry skull. The lengths of the markers were measured with a high-precision caliper. The skull and the template were scanned in each of the four CBCT devices (1. Gendex GXCB-500; 2. Sirona Galileos Comfort; 3. Sirona Orthophos XG 3D; 4. Carestream CS 9300) 19 times (10 scans without moving the skull, and 9 scans with repeated repositioning of the skull in the device, according to the manufacturers' instructions). A 3D implant-planning program was used to measure the lengths of the six markers digitally. Actual and digital measurements were compared to determine device-specific errors. The repositioning of the skull examined the reproducibility of the CBCT devices. Linear measurements were analyzed statistically (P 0.05). Results: Mean deviations without moving the skull (vertical/sagittal/transverse) for device 1 were 0.023 mm/0.000 mm/0.025 mm (0.07%/0.19%/0.24%), for device 2 were 0.410 mm/0.115 mm/0.080 mm (-1.75%/0.32%/0.88%), for device 3 were -0.665 mm/-0.215 mm/-0.675 mm (-2.71%/-1.82%/-4.42%), and for device 4 were -0.045 mm/-0.135 mm/-0.410 mm (-0.45%/-1.54%/-2.57%). The overall mean deviation for device 1 was 0.028 mm (0.16%), for device 2 was 0.072 mm (-0.95%), for device 3 was 0.518 mm (-2.97%), and for device 4 was -0.197 mm (-1.53%). The mean deviation after repositioning for device 1 was 0.004 mm (-0.65%), for device 2 was -0.250 mm (0.95%), for device 3 was 0.496 mm (-2.66%), and for device 4 was -0.265 mm (-1.92%). Thus, apart from device 3, the deviations increased. Conclusion: Deviations from the actual measurements were detected with each device. Therefore, respecting safety distances when placing implants is crucial. Palabras clave: implant planning, 3D planning, CBCT, 3D imaging, radiographic reproducibility, radiographic accuracy