Radioisotopes find very important applications in various sectors of economic significance and their production is an important activity of many national programmes. Some deterministic codes such as ALICE ASH 1.0 and TALYS 1.0 are extensively used to calculate the yield of a radioisotope via numerical integral over the calculated cross-sections. MCNPX 2.6 stochastic code is more interesting among the other Monte Carlo-based computational codes for accessibility of different intranuclear cascade physical models to calculate the yield using experiment-based cross-sections. A benchmark study has been proposed to determine the codes' uncertainty in such calculations. ${}^{109}$Cd, ${}^{86}$Y and ${}^{85}$Sr production yields by proton irradiation of silver, rubidium chloride and strontium carbonate targets are studied. $^{109}$Cd, $^{86}$Y and $^{85}$Sr cross-sections are calculated using ALICE ASH 1.0 and TALYS 1.0 codes. The evaluated yields are compared with the experimental yields. The targets are modelled using MCNPX 2.6 code. The production yields are calculated using the available physical models of the code. The study shows acceptable relative discrepancies between theoretical and experimental results. Minimum relative discrepancy between experimental and theoretical yields is achievable using ISABEL intranuclear model in most of the targets simulated by MCNPX 2.6. The stochastic code utilization can be suggested for calculating $^{109}$Cd, $^{86}$Y and $^{85}$Sr production yields. It results in more valid data than TALYS 1.0 and ALICE ASH 1.0 in noticeably less average relative discrepancies.