The aim of this animal study is to analyze bone remodeling around platform-switching (PS) implants with and without a machined (MACH) collar placed at different levels in relation to the crestal bone in a canine model. All mandibular premolars and first molars were extracted in five dogs. After 6 months, grit-blasted acid-etched (GBAE) PS implants with and without a MACH neck were randomly inserted in each hemimandible, positioning the implant-abutment interface in either a supracrestal (+1.5 mm), equicrestal, or subcrestal (-1.5 mm) position, and healing abutments were connected. Implant abutments were dis/reconnected at 12, 14, 16, and 18 weeks after implant placement. After 6 months of healing, animals were sacrificed. Clinical parameters and periapical radiographs were registered on the day of implant placement, at 2 months, at every abutment dis/reconnection, and at sacrifice. Crestal bone changes were calculated and defined as the primary outcome variable. When crestal bone changes from implant placement to 6 months were compared between MACH and GBAE groups, no significant differences were encountered except for implants placed in an equicrestal position (P = 0.04). However, multivariable regression analysis revealed no significant differences between MACH and GBAE implants placed in a supracrestal (β = -0.08; P = 0.45), equicrestal (β = -0.05; P = 0.50), or subcrestal (β = -0.13; P = 0.19) position. Surface treatment of the implant neck had no significant influence on crestal bone changes around PS implants with and without a MACH collar.