Purpose: To compare the accuracy of two digital workflows for producing resin patterns to be cast into metal frameworks compared to an identical framework manufactured conventionally from a wax pattern. Materials and Methods: Nine casts were duplicated from a maxillary master cast of a partially edentulous arch. Their accuracy was determined by measuring the same points in two and three dimensions using a reflex microscope, which was also used to measure all frameworks to an accuracy of 4 μm. The same design was used throughout. Three casts were used to make a framework conventionally from an invested wax pattern. Six casts were scanned, and a digital pattern created. Three patterns were milled from a resin block, and three were 3D printed with resin. Then each pattern was cast. Results: The sample size precluded direct statistical conclusions, but no significant differences were found. Duplicate models showed minimal differences compared to the master cast. All patterns and all frameworks showed some level of difference compared to the master cast, but no differences were greater than those reported in the literature as being clinically acceptable. The maximum overall discrepancy between the cast frameworks was 0.64 mm, and at the rest seats was 0.262 mm. Conclusion: Within the limitations of this study, given the very small actual differences both within and between the groups of the three different workflows, the use of digitally produced resin patterns prior to their being cast as metal frameworks is both feasible and well within the accepted limits for clinical acceptability.