Torque failure for endosseous implants has been used as a biomechanical measure of anchorage, though the significance of such data is not known. A practical understanding of the resistance to torque failure of implant-tissue interfaces at stage 2 surgery would help in assigning torque levels for implant abutment-screw fastening. The purpose of this study was to measure torque failure levels of commercially pure (CP) titanium, Ti-6Al-4V, hydroxyapatite-coated (HA-coated) screw-shaped implants placed into the maxillae and mandibles of baboons. Implants identical in size were placed into the edentulous posterior maxillae and mandibles of six female baboons (n =37 each group) using a standardized surgical protocol. Reverse-torque data were collected at postinsertion time intervals ranging from 3 to 4 months using a counterclockwise torque driver and the data were analyzed (repeated measures ANOVA) for torque differences related to time, biomaterial, and jaw. The HA-coated implants exhibited significantly greater torque-removal values compared to both metallic implants (HA: 186.0 Ncm [50.1]; Ti-6Al-4V: 78.6 Ncm [18.1]; CP Ti 74.0 Ncm [24.4]). Analysis of torque interactions with jaw showed no significant difference; however, the mandible was found to be greater than the maxilla in torque resistance for all groups tested. Understanding the risks in inferring animal data to human application, the clinical implications of these data suggest that the recommended torque level of 35 Ncm for abutment fastening may provide a margin of safety for most implants of similar design and material as used in this study. The range of data suggests that 1 in 20 metallic implants may fall below this torque-fastening level. Palabras clave: baboon jaws, commercially pure titanium, hydroxyapatite-coated implants, Ti-6Al-4V implants, torque failure