Purpose: The effects of implant angulation, impression material, and variation in width of the arch curvature on transfer models were evaluated. Materials and Methods: Three edentulous maxillary epoxy resin models were fabricated, and six internal-connection implant analogs were placed in different locations and different angulations in each model. In the first model, implants were positioned in the canine, first premolar, and first molar regions, and all analogs were positioned parallel to each other and perpendicular to the horizontal crestal plane (parallel model). In the second model, analogs were positioned in same regions (canine, first premolar, and first molar), but three of them were positioned with 10-degree buccal angulations (versus the horizontal crestal plane) (angular model). In the third model, analogs were inserted in the lateral incisor, canine, and second molar regions and parallel to each other (wide-arch model). Eighteen impressions of each model were made with each of the three materials-condensation silicone, polyvinyl siloxane, and polyether-and impressions were poured and kept at room temperature for 24 hours. They were then observed under a toolmaker's microscope, with epoxy resin models of each group used as references. Distance deformations between implants in each model in the x- and y-axes were recorded separately. Implant angulation deformations were recorded in the x-z plane. Statistical evaluations were performed with analysis of variance and the least significant difference post hoc test. Results: Angular model measurements showed the greatest deformation values (P < .05). All impression materials showed deformation, and the polyether impression models showed statistically significantly less deformation in angular measurements (P < .05). Conclusions: The models with implants placed parallel to each other exhibited greater accuracy than a model with implants placed at angles to each other.