Six cadaver forelimbs were imaged in two high-field magnetic resonance (MR) systems and one low-field MR system following the creation of osteochondral defects on the palmar distal aspect of the third metacarpal bone. The following sequences were performed using all three systems: proton density (PD) turbo spin echo, T2* gradient echo (GRE), T2-weighted fast spin echo, and short tau inversion recovery. In addition, 3D T1 GRE sagittal standard and motion insensitive sequences were obtained using the low-field system. PD fat saturated and 3D T1-weighted spoiled GRE images with and without fat suppression were acquired with the high-field systems. Lesions were measured and assigned a confidence score. The images obtained using high-field systems (1.0 and 1.5 T) more accurately represented the osteochondral defects when compared with low-field system (0.27 T) images. The largest difference was observed when evaluating articular cartilage defects, which were not identified on the low-field images. Sequence selection affected the appearance of the lesions. On all systems the turbo and fast spin echo sequences more accurately represented the lesion size and shape when compared with the GRE sequences. The T1 GRE sequence is the only sequence that appears to allow visualization of the articular cartilage on the low-field images, but is limited in providing adequate cartilage visualization. Confidence scores were greater on the high-field systems when compared with the low-field system.