Sensorineural hearing loss (SNHL) is the most common sensory deficit in the world, caused by damage to cellular structures within the inner ear, or cochlea. Visualization of the cellular pathology underlying different types of SNHL has been difficult due to the small size of the cochlea, its complex three-dimensional structure, and embedded location within the temporal bone. Micro-optical coherence tomography (µOCT) is a recently-developed cross-sectional imaging technology that can obtain images with sufficient detail to elucidate specific aetiologies of SNHL. In this work, we developed a high resolution, ultra-small-diameter, flexible probe for imaging the human cochlea in situ. The 500 µm diameter, circumferential scanning µOCT imaging probe contains self-imaging wavefront division optics that provide maximal lateral resolution of 2.5 µm and better than 5 µm resolution over an extended depth of focus of 1 mm in air. Using a supercontinuum light source with a 300 nm bandwidth and a common path interferometery configuration, axial resolution is 1.9 µm in air. Images of 3D-intact cochleae extracted from human cadavers were acquired with the µOCT probe in situ; these images demonstrate the system’s ability to visualize the entire cross-section of the scala tympani, in addition to cellular structures in the cochlea’s sensory epithelium, the organ of Corti and bundles of auditory nerves. These results suggest that this new device has the potential to facilitate personalized diagnosis and therapy for SNHL.