Intracochlear injection through the round window membrane (RWM) has been proposed to overcome imprecise drug delivery into the inner ear. Using a novel ultrasharp microneedle, we compared the perilymphatic dexamethasone (DEX) concentration achieved after intratympanic vs. intracochlear injection at two different time points and assessed its safety in guinea pigs. For this purpose, DEX sodium phosphate (10 mg/mL) was administered either in the right middle ear space via continuous intratympanic injection or in the right scala tympani of the cochlea with microneedle-mediated injection (1 µL at 1 µL/min) across the RWM. Both groups were evaluated at 1-hour or 3-hour time points. Perilymph from both cochleae was sampled for liquid chromatography-mass spectrometry, and bilateral cochleae were harvested for immunofluorescence. Eighteen guinea pigs were included. The mean DEX concentration was higher in the intracochlear delivery group than in the intratympanic delivery group at 1-hour time point (mean difference 67,863 ng/mL, 95% CI (8,352-127,374 ng/mL), p = 0.03). No difference was found at 3-hour time point. In every animal on both cochleae, no disruption in hair and supportive cells of the organ of Corti and utricle was observed. Significant middle ear inflammation was observed with the intratympanic delivery method compared to intracochlear. In conclusion, microneedle-mediated intracochlear injection achieves higher perilymphatic DEX concentration than the intratympanic route by a factor of 7 while preserving the cochlear architecture and inducing significantly less middle ear inflammation. In this new era of inner ear therapeutics, the potential for translational application is tangible and promising.