Acoustic information in the auditory system is captured in space-time patterns of neural spiking. Ideally, cochlear implants should convey information contained in sound to the central auditory nervous system through electrically elicited neural activation patterns that would result in a sound sensation indistinguishable from that provided through acoustic auditory stimulation. Computational models of nerve fiber firings elicited by acoustic or electrical stimulation of the cochlea may be used to investigate the characteristics of electrically stimulated neurons, the resulting neural spike train patterns, and how these encode acoustic information. This chapter commences by considering the response of single nerve fibers captured by models based on or related to the classical Hodgkin-Huxley nerve fiber model and proceeds to model specializations specific to electrical stimulation of nerve fibers. Next, end-to-end models, which capture processing from the acoustic input to the neural activity output of the cochlea, are discussed. The latter is in the form of a neurogram that reflects neural activity across the entire cochlea. How this population response varies for different peripheral input is considered here for normal hearing, a commercial speech processing algorithm, and a custom speech processing algorithm. This demonstrates how neurograms may be used to investigate differences between speech processors. Proceeding from this, performance measures derived from neurograms are discussed. The neural activity output from the cochlea is processed through several stages before eliciting a perception. There is no unified approach to modeling this bridge between stimulus and perception. A cochlear nucleus model to predict pitch perception is presented and modeling of processing beyond the cochlear nucleus is discussed. Examples of AI approaches to modeling of psychoacoustic measures and speech intelligibility are discussed.