The purpose of this study was to explore the operator performance of the fabrication of digital orthodontic setups integrated into cone beam computed tomography (CBCT) scans. Fifteen patients who underwent a combined orthodontic–orthognathic surgical treatment were included. The pre-treatment digital dental models and CBCT scans were fused, and four operators made virtual setups twice for all patients. Differences between the virtual setups were calculated by recording tooth crown movement from the pre-treatment model to the virtual setup. To examine performance, Pearson’s correlation coefficients, duplicate measurement errors, and inter-operator differences were calculated. For intra-operator performance, correlation values varied among tooth types, with mean correlation values from 0.66 to 0.83 for the maxilla and 0.70 to 0.83 for the mandible. For inter-operator performance, mean correlation values varied from 0.40 to 0.87 for the maxilla and from 0.44 to 0.80 for the mandible. Rotational mean differences exceeded the range of clinical acceptance (>2 degrees) at 18% for the maxilla and 20.8% for the mandible, and translational mean differences exceeded the range of clinical acceptance (0.6 mm) at 9.7% and 26% for the maxilla and mandible, respectively. The intra- and inter-operator performance of digital orthodontic setup construction for virtual three-dimensional orthognathic planning shows significant errors.