Researchers have attempted to model the magnetomechanical behavior of magnetic shape memory alloys (MSMAs) for over a decade, but all of the models developed to date have only been validated against experimental data generated under two-dimensional loading conditions. As efforts have been underway to develop models able to predict the most general (i.e. 3D) loading conditions for the material, there is a need for experimental data to support the calibration and validation of these models. This paper presents magnetomechanical data from experiments where a MSMA specimen whose microstructure accommodates three martensite variants is subjected to three-dimensional magneto-mechanical loading. To the best of our knowledge, all prior experimental investigations on MSMAhave been performed onsamples accommodating two martensite variants and exposed to two-dimensional magnetomechanical loads. The experimental results from the 3D loading of the three variant MSMA specimen are used to calibrate and validate a 3D model developed by this group [LaMaster et al. (2014)]. This model assumes that three martensite variants coexist in the material. The LaMaster et al. model captures the general trends seen in the experimental data, but does not predict the data with a high degree of accuracy. Possible reasons for the mismatch between experimental data and model predictions are discussed.