Quadrupedal locomotion arises from a complex interplay between spinal neuronal circuits, descending brain signals, musculoskeletal interactions, and sensory feedback, enabling adaptive speed and terrain dependent gait transitions. However, the underlying neural mechanisms involved in limb coordination are poorly understood. Here, we present a 3D closed-loop neuromechanical model of a mouse and use it to explore interactions between the spinal circuitry and afferent feedback and their role in interlimb coordination during quadrupedal locomotion. The spinal circuit model includes four rhythm generators, each controlling one limb, that define the locomotor frequency and flexor–extensor alternation. The rhythm generators control pattern formation circuits that generate muscle synergies and create muscle-specific activation patterns. Commissural interneurons mediate (fore-hind and left-right) interlimb coordination. Afferent feedback (muscle spindle Ia and II, Golgi tendon Ib, and cutaneous) is connected to the rhythm generators (affecting timing of phase-transitions), to the pattern formation circuits (affecting muscle synergies), and directly to motoneurons and premotor interneurons, forming basic reflex circuits. Motoneurons activate the muscles to generate locomotor behaviors. Using evolutionary strategies, the model was optimized to successfully locomote over flat, sloped, and uneven terrains. Interactions of sensory feedback with spinal neural circuitry allow for locomotor adaptation to different environmental conditions. The model suggests that the role of sensory feedback signals in control of locomotion depends on the speed and gait; this dependence is subject to further investigation. Overall, the neuromechanical mouse model provides an open-source tool that can be used as a testbed to study complex locomotor behaviors in 3D environments.