Anisotropic yield models are widely used to meet the design and safety margins of engineering components made from metallic alloys under complex forming processes and other multi-axial loading conditions. Accurately calibrating the material parameters in these models is essential for ensuring realistic predictions of the directional deformation behavior of polycrystalline metals. This study compares the effects of different considerations in crystal-plasticity finite element simulations for the conduction of virtual experiments intended for the calibration of an anisotropic yield function for an aluminum 6061-T6 sheet. Micromechanics models informed by electron backscatter diffraction (EBSD) require significant computational cost, while increasingly complex models require additional time to set up and verify their accuracy. This study is intended to determine which considerations that increase the complexity of the rate-dependent visco-plastic crystal plasticity finite element model also make a significant improvement to the accuracy of the simulations. It explores the effects of the hardening model formulation (slip-based compared to dislocation-density-based models), assumed material rate-dependence, the homogenization method for calculating macroscopic behavior from the micro-scale model, the modeling of precipitates in the microstructure, the number of finite elements to the number of grains in the model, and the effect of grain morphology on the predictions of the material behavior. The predictions from the computational models are compared to one another and to the results of a separate experimental campaign reported in earlier work. This work explores the viability of multiscale modeling for the calibration of advanced plasticity models for a 6061-T6 aluminum and which complexifying assumptions in the model make a significant impact on the accuracy of the results.