Monte Carlo simulations provide accurate results for the neutronic response of the system under consideration if modeling is performed appropriately since it has great influence on the results. Sensitivity analysis of modeling approaches for geometry and fissile material composition distributions in the reactor core was performed by taking ITU TRIGA Mark II Research Reactor into consideration. The method of defining fuel element positions in the core by using circular or hexagonal lattice was considered as one case and three different methods of lumping material compositions in the fuel elements was considered as another case since these approaches are used by deterministic codes hence the accuracy of deterministic codes were also investigated. The validation study showed that both MCNP and Serpent Monte Carlo codes resulted in good agreement with the experimental data. It was observed that the handling of fuel composition in different ways did not influence the results significantly (up to 11.1 cents in reactivity). However, the influence of fuel arrangement is more pronounced (deviation in reactivity calculations is around 1$). These deviations at the results may affect the nuclear safety conclusion of reactors having small shutdown margins. It was also concluded that users of the deterministic codes should be aware of the fact that the simplifications in geometry and fuel composition in the core will result in significant deviation from the reality.