The purpose of this work is the modeling of the deformation behavior and orientation gradient development in a highly anisotropic thin metal sheet and comparison with experiment. This sheet consists of a single layer of "large" Fe-3%Si grains exhibiting a coarse texture. Since such materials are highly heterogeneous, they are modeled by combining single-crystal plasticity for each grain with the finite-element method for the grain morphology and specimen as a whole. The single-crystal model is rate-dependent, accounts for (local) dissipative hardening effects, and has been identified with the help of single-crystal data. In previous work Klusemann et al. (2012b), model predictions for the evolution of the specimen geometry and grain morphology during tension loading to large deformation have been shown to agree reasonably well with the corresponding experimental results of Henning and Vehoff (2005). In the current work, model predictions for the development of orientation gradients in the specimen under different modeling assumptions (e.g., active glide-system family) are compared with EBSD-based experimental results of Henning and Vehoff (2005). Model predictions for the development of geometrically necessary dislocations are also discussed. As well, additional measures of local orientation evolution such as reorientation are examined and compared with the orientation gradient picture. In addition, we examine the effect of additional grain boundary strengthening related to grain boundary misorientation and grain size and the effect of additional GND-based kinematic hardening.