In evolutionary biomechanics, musculoskeletal computer models of extant and extinct taxa are often used to estimate joint range of motion (ROM) and muscle moment arms (MMAs), two parameters which form the basis of functional inferences. However, relatively few experimental studies have been performed to validate model outputs. Previously, we built a model of the short-beaked echidna (Tachyglossus aculeatus) forelimb using a traditional modelling workflow, and in this study we evaluate its behaviour and outputs using experimental data. The echidna is an unusual animal representing an edge-case for model validation: it uses a unique form of sprawling locomotion, and possesses a suite of derived anatomical features, in addition to other features reminiscent of extinct early relatives of mammals. Here we use diffusible iodine-based contrast-enhanced computed tomography (diceCT) alongside digital and traditional dissection to evaluate muscle attachments, modelled muscle paths, and the effects of model alterations on the MMA outputs. We use X-ray Reconstruction of Moving Morphology (XROMM) to compare ex vivo joint ROM to model estimates based on osteological limits predicted via single-axis rotation, and to calculate experimental MMAs from implanted muscles using a novel geometric method. We also add additional levels of model detail, in the form of muscle architecture, to evaluate how muscle torque might alter the inferences made from MMAs alone, as is typical in evolutionary studies. Our study identifies several key findings that can be applied to future models. 1) A light-touch approach to model building can generate reasonably accurate muscle paths, and small alterations in attachment site seem to have minimal effects on model output. 2) Simultaneous movement through multiple degrees of freedom, including rotations and translation at joints, are necessary to ensure full joint ROM is captured; however, single-axis ROM can provide a reasonable approximation of mobility depending on the modelling objectives. 3) Our geometric method of calculating MMAs is consistent with model-predicted MMAs calculated via partial velocity, and is a potentially useful tool for others to create and validate musculoskeletal models. 4) Inclusion of muscle architecture data can change some functional inferences, but in many cases reinforced conclusions based on MMA alone.