In a stop signal paradigm to investigate the control of human saccades subjects were instructed to make a saccade to a visual target appearing suddenly l5 degrees to the left or to the right of the fixation point. In 25% of the trials an auditory stop signal was presented after a variable delay that required the subject to inhibit the saccade. The stop signal was presented randomly at the target position, at the opposite side, or at fixation. Using different estimation techniques the average time needed to inhibit a saccade (stop signal processing time, or SSPT) was estimated on the basis of the race model. The SSPT estimates ranging from 50 to 100 ms (depending on subject) are shorter than those from previous studies with visual stop signals. Position of the auditory stop signal did not show an effect on countermanding effectiveness. We found saccadic response times consistent with the race model predictions for two subjects, while a third subject showed small but consistent violations. Moreover, all subjects showed a tendency towards hypometric saccades for responses that could not be inhibited. These findings are discussed with respect to recent neurophysiological results.