OBJECTIVE: The objective of this study was to determine how the presentation of unprocessed speech, either ipsilaterally (to simulate electro-acoustic stimulation, EAS) or contralaterally (to simulate bimodal stimulation), alongside vocoder-processed speech affects the efficiency of spoken word processing. METHOD: Gated word recognition was performed under four listening conditions: full-spectrum speech, vocoder-processed speech, electro-acoustic stimulation (EAS), and bimodal stimulation. In the EAS condition, low-frequency unprocessed speech and high-frequency vocoder-processed speech were presented to the same ear, while in the bimodal condition, full-spectrum speech was presented to one ear and vocoder-processed speech to the other. RESULTS: Listeners identified target words accurately with just over half of the word duration in the full-spectrum condition, whereas nearly the entire word was required for correct identification with vocoder-processed speech. Combining full-spectrum speech and vocoder-processed speech, whether ipsilaterally or contralaterally, led to significant improvements in gated word recognition. Full-spectrum speech yielded the best-gated word recognition performance, followed by bimodal stimulation, with intermediate results from EAS and the lowest performance from the vocoder condition. CONCLUSION: Inherent limitations in CI spectral resolution can impair spoken word processing. Adding acoustic stimulation can improve gated word recognition performance. Bimodal stimulation significantly enhances lexical processing timing compared to EAS in one ear.