This study was conducted to evaluate the accuracy of 3D-printed surgical guides before and after sterilization in a steam sterilizer. A test-model incorporating three implant replicas was customized. A total of forty guides were printed from five printable resins. A group made from a self-curing composite served as control group. The guides were checked for fit. Vertical discrepancies between the model and guides were measured at standardized points at a load of 500 g (P1). The guides were connected to implant replicas and scanned, and their angles were digitally measured. The specimens were sterilized in a steam sterilizer at 121 °C for 20 min at 2 bar pressure. Vertical discrepancies (P2) and angulations were remeasured. Additionally, the specimens were repositioned with an increased load, and measurements were repeated (P3). All specimens were repositionable after sterilization. The smallest variation in discrepancy at a 500 g load was 428 µm, whereas the greatest was 1487 µm. Under an increased force, the smallest change was 94 µm, while the greatest was 260 µm. The level of significance α = 0.05 (95% confidence interval) was set for all tests. The variation in the measured angles was not statistically significant (Kruskal-Wallis's test, > 0.05). The accuracy was affected by the material and sterilization, but it was clinically acceptable when an increased load was applied during repositioning.