During cochlear implantation, electrode array translocation and trauma should be avoided to preserve residual hearing. The aim of our study was to evaluate the effect of physical parameters of the array on residual hearing and cochlear structures during insertion. Three array prototypes with different stiffnesses or external diameters were implanted in normal hearing guinea pigs via a motorized insertion tool carried on a robot-based arm, and insertion forces were recorded. Array prototypes 0.4 and 0.4R had 0.4 mm external diameter and prototype 0.3 had 0.3 mm external diameter. The axial stiffness was set to 1 for the 0.4 prototype and the stiffnesses of the 0.4R and 0.3 prototypes were calculated from this as 6.8 and 0.8 (relative units), respectively. Hearing was assessed preoperatively by the auditory brainstem response (ABR), and then at day 7 and day 30 post-implantation. A study of the macroscopic anatomy was performed on cochleae harvested at day 30 to examine the scala location of the array. At day 7, guinea pigs implanted with the 0.4R array had significantly poorer hearing results than those implanted with the 0.3 array (26±17.7, 44±23.4, 33±20.5 dB, n = 7, vs 5±8.7, 1±11.6, 12±11.5 dB, n = 6, mean±SEM, respectively, at 8, 16 and 24 kHz, p<0.01) or those implanted with the 0.4 array (44±23.4 dB, n = 7, vs 28±21.7 dB, n = 7, at 16 kHz, p<0.05). Hearing remained stable from day 7 to day 30. The maximal peak of insertion force was higher with the 0.4R array than with the 0.3 array (56±23.8 mN, n = 7, vs 26±8.7 mN, n = 6). Observation of the cochleae showed that an incorrectly positioned electrode array or fibrosis were associated with hearing loss ≥40 dB (at 16 kHz). An optimal position in the scala tympani with a flexible and thin array and prevention of fibrosis should be the primary objectives to preserve hearing during cochlear implantation.