To compare the accuracy of implant master casts fabricated using Robocast Technology (Biomet 3i) with that of master casts fabricated using traditional transfer (closed-tray) and pick-up (open-tray) techniques. A stereolithographic replica of a Kennedy Class I human mandible was fabricated for use as the master model. Implants were placed into both posterior quadrants (both second premolars and second molars) and set parallel (P) on one side and divergent (nonparallel, NP) on the opposite side. Impressions were made of the master model (patient replica model) with Encode Healing Abutments, open-tray, and closed-tray impression copings. Identical metallic spheres were placed onto each implant analog in the stone master casts, and the casts/spheres were scanned using a digital scanner. Measurements were made between the center points of the spheres and compared to the master model. Data were divided into P, NP, and individual sites, and the differences were analyzed statistically (α = .05). Encode master casts were less accurate than the open-tray casts in NP sites. Encode master casts were less accurate than the open-tray and closed-tray casts in P sites. NP sites demonstrated less accuracy than P sites within the Encode group. Encode master casts were less accurate than the open- and closed-tray casts at the mandibular right second premolar site. The mandibular left second premolar was less accurate than the mandibular right second molar in the Encode group. Within the limitations of this lab-based study and analysis, the Encode technique resulted in master casts that were less accurate than master casts made from traditional open- and closed-tray impression techniques. Further research is necessary before specific clinical judgments can be made.