Constructing musculoskeletal models of extinct vertebrates requires subjective assumptions about soft tissue parameters rarely preserved in the fossil record. Despite these necessary assumptions about fundamental input values, paleobiologists rarely perform objective tests of best-estimate models before reaching conclusions based on predicted results. The extent to which lack of knowledge on soft tissue anatomy limits the accuracy of running speed estimates of extinct dinosaurs is therefore poorly understood. In this study, a sensitivity analysis is performed on an evolutionary robotics model of the non-avian theropod dinosaur Allosaurus, used previously to estimate maximum running speed in this extinct animal. A range of muscle parameters were varied over the range observed in extant vertebrates, whereas mass-related parameters were altered across the range of published estimates for Allosaurus. Muscle parameters have a linear relationship with maximum running speed, whereas surprisingly total body mass and torso center of mass have little effect. Muscle force values produced the greatest range in predicted running speeds (4.5–10.7 m/s) and stride lengths (4–5.8 m) in the sensitivity analysis, equating to 65.9% and 30.7% variation about the original ‘best-estimate’ prediction, a relatively high potential margin of error. These results highlight the importance of sensitivity analyses in biomechanical modeling of extinct taxa, particularly where values for soft tissues parameters are not tightly constrained. The current range in plausible values for soft tissue properties makes a robust quantitative assessment of behavioral ecology and species interactions in dinosaurian communities extremely difficult.