Observers were tested in perceptual category-learning experiments in which they were instructed to make classification decisions as rapidly as possible without making errors. Nosofsky and Palmeri's (1997) exemplar-based random walk (EBRW) model of speeded classification was tested on its ability to fit the classification response times and accuracies. In Experiment 1, the authors demonstrated that the EBRW provided good quantitative fits to the response times and accuracies associated with individual objects as a function of their locations in a multidimensional similarity space and as a function of practice in the task. In Experiment 2, the authors demonstrated that individual-object familiarity can play an important role in speeded classification. Unfamiliar objects were classified more slowly than familiar ones, even when the unfamiliar objects were not completely novel. The EBRW provided a reasonably good quantitative account of the familiarity effects on the classification response times.