Abstract Objective In total hip arthroplasty (THA), surgeons attempt to achieve a physiological antetorsion. However, postoperative antetorsion of the femoral stem is known to show large variabilities. The purpose of this study was to assess whether postoperative antetorsion is influenced by stem design or cementation. Materials and methods This retrospective study included 227 patients with a hip prosthesis with five different stem designs (S1: short curved, S2 and S3: standard straight, S4: standard straight collared, S5: cemented straight), who had metal suppressed 1.5T-MRI of the hip between February 2015 and October 2019. Measurement of femoral antetorsion was done independently by two fellowship-trained radiologists on axial images by measuring the angle between the long axis of the femoral neck and the posterior condylar tangent of the knee. Measured angles in the different groups were compared using the t test for independent samples. Results The cementless collared stem S4 showed the highest antetorsion with 18.1° (± 10.5°; range –10°–45°), which was significantly higher than the antetorsion of the collarless S3 with 13.3° (± 8.4°; − 4°–29°) and the cemented S5 with 12.7° (± 7.7°; − 3°–27°) with p = 0.012 and p = 0.007, respectively. S1 and S2 showed an antetorsion of 14.8° (± 10.0°; 1°–37°) and 14.1° (± 12.2°; − 20°–41°). The torsional variability of the cementless stems (S1–4) was significantly higher compared with that of the cemented S5 with a combined standard deviation of 10.5° and 7.7° (p = 0.019). Conclusion Prosthesis design impacts the postoperative femoral antetorsion, with the cementless collared stem showing the highest antetorsion. Cemented stems demonstrated significantly lower variability, suggesting the lowest rate of inadvertent malrotation.