Finite-element analysis (FEA) is widely-used in paleontology, but accurately modeling extinct taxa is difficult. In particular, properties of sutural tissues are poorly characterized, and modeling them is time-consuming. Validation studies on extant taxa indicate suture inclusion may have little impact on overall strain patterns. However, these have mostly been performed on mammals, whose box-like skulls are unlike those of sauropsids, which include many long, overlapping sutures. We present FEA of the skull of the Early Jurassic theropod Megapnosaurus kayentakatae which retains many unfused sutures. These include a loose premaxilla-maxilla joint, hypothesized to have either modified local strain transmission, or facilitated passive kinesis, during biting. The skull was restored in Avizo, with jaw muscles reconstructed from osteological correlates. FEA models were built in Strand7, with material properties based on extant analogs. Models were solved with and without sutures for unilateral and bilateral bites along the toothrow. Model comparison shows that including sutures yields significant differences in absolute magnitudes of stress and strain in the skull of Megapnosaurus. Further, the loose premaxilla-maxilla articulation redistributes local strains, but high bending stresses in the nasals prevent kinesis. Nonetheless, global comparison reveals that sutures have little impact on overall patterns of relative stress and strain distribution. Consequently, if relative performance is of interest, excluding cranial sutures is a reasonable compromise, but fine-scale information will be lost. This highlights the importance of considering research questions thoroughly before designing model complexity.