The purpose of this study was to evaluate the accuracy of fitting of an implant supported screw-retained bar made on definitive casts produced by 4 different dental stone products. The dental stones tested were QuickRock (Protechno), FujiRock (GC), Jade Stone (Whip Mix) and Moldasynt (Heraeus). Three external hexagon implants were placed in a polyoxymethylene block. Definitive impressions were made using monophase high viscosity polyvinylsiloxane in combination with custom trays. Then, definitive models from the different types of dental stones were fabricated. Three castable cylinders with a machined non-enganging base were cast and connected with a very small quantity of PMMA to a cast bar, which was used to verify the marginal discrepancies between the abutments and the prosthetic platforms of the implants. For that purpose special software and a camera mounted on an optical microscope were used. The gap was measured by taking 10 measurements on each abutment, after the Sheffield test was applied. Twelve definitive casts were fabricated for each gypsum product and 40 measurements were performed for each cast. Mean, minimum, and maximum values were calculated. The Shapiro-Wilk test of normality was performed. Mann-Whitney test (P<.06) was used for the statistical analysis of the measurements. The non-parametric Kruskal-Wallis test revealed a statistically significant effect of the stone factor on the marginal discrepancy for all Sheffield test combinations: 1. Abutment 2 when screw was fastened on abutment 1 (?2=3, df=35.33, P<0.01), 2. Abutment 3 when the screw was fastened on abutment 1 (?2=3, df=37.74, P<0.01), 3. Abutment 1 when the screw was fastened on abutment 3 (?2=3, df=39.79, P<0.01), 4. Abutment 2 when the screw was fastened on abutment 3 (?2=3, df=37.26, P<0.01). A significant correlation exists between marginal discrepancy and different dental gypsum products used for the fabrication of definitive casts for implant supported bars. The smallest marginal discrepancy was noted on implant supported bars fabricated on definitive casts made by Type III mounting stone. The biggest marginal discrepancy was noted on implant supported bars fabricated on definitive casts made by Type V dental stone. The marginal discrepancies presented on implant supported bars fabricated on definitive casts made by two types of Type IV dental stone were not significantly different.