Purpose: To evaluate the efficacy of short (5 or 6 mm long) dental implants vs 10 mm or longer implants placed in crestally lifted sinuses. Materials and methods: Twenty partially edentulous patients with 5 to 7 mm of residual crestal height and at least 7 mm thickness below the maxillary sinuses as measured on computed tomography scans were randomised according to a parallel group design to receive either one to two 5 or 6 mm long implants (10 patients) or 10 mm long implants (10 patients) after crestal sinus lifting and grafting with anorganic bovine bone. Implants were left to heal submerged for 4 months and were loaded with reinforced acrylic provisional prostheses, replaced, after 4 months, by definitive provisionally cemented or screw-retained metal-ceramic or metal-composite prostheses. Outcome measures were: prosthesis and implant failures, any complications, radiographic peri-implant marginal bone level changes, and patient's satisfaction assessed by blinded assessors, when possible. All patients were followed up to 3 years after loading. Results: Two patients from the augmented group dropped out after the 1-year follow-up. No implant or prosthesis failure occurred. One short implant patient was affected by two complications vs one complication in the long implant group. There was no difference for patients experiencing complications between the two groups (difference in proportions = -0.03 %; 95%CI: -0.32 to 0.27; P = 1.000). Short implants lost 0.89 ± 0.25 mm and long implants lost 1.08 ± 0.29 mm of peri-implant marginal bone 3 years after loading, the difference between the two groups being not statistically significant (difference = -0.19 mm; 95%CI: -0.46 to 0.09; P = 0.165). All patients were fully or partially satisfied with function and fully satisfied with aesthetics. Conclusions: Both techniques achieved excellent results and no differences were observed between prostheses supported by one to two 5 or 6 mm long implants vs 10 mm long in posterior atrophic maxillae up to 3 years after loading; therefore it is up to clinicians to decide which procedure to use, although longer follow-ups with larger patient populations are needed to better understand if one of these procedures could be more effective in the long-term.