Binocular rivalry is a widely used tool in sensory and cognitive neuroscience to investigate different aspects of vision and cognition. The dynamics of binocular rivalry (e.g., duration of perceptual dominance phases and mixed percept proportions) differ across individuals; based on rivalry dynamics, it is also possible to calculate an index of ocular dominance (by comparing the perceptual dominance of the images in the two eyes). In this study, we investigated the reliability of binocular rivalry dynamics using different methods for dichoptic stimulation and different rivalry stimuli. For the three main indices we defined (ocular dominance, phase durations and mixed percept proportions), we found a high test–retest reliability across sessions. Moreover, the test–retest reliability of the ocular dominance index was predictable from its internal consistency, supporting its stability over time. Phase durations and mixed percept proportions, in contrast, had worse test–retest reliability than expected based on internal consistency, indicating that these parameters are susceptible to state-dependent changes. Our results support the use of the ocular dominance index and binocular rivalry in the measurement of sensory eye dominance and its plasticity, but advise caution when investigating the association between phase durations or mixed percepts and stable characteristics like psychological traits or disorders.