Current clinical treatments for neuropathic pain include amitriptyline, a tricyclic antidepressant with mixed pharmacology that is also clinically reported to impair cognitive performance; and gabapentin, a compound that selectively interacts with alpha2delta-1 calcium channel subunits. Since few assessments of cognitive performance have been made in non-human primates with these marketed treatments, the purpose of this study was to determine their relative abilities to alter working memory as measured in mature macaques in their performance of a delayed matching-to-sample task. Four delay intervals of increasing duration provided increasing impairment in task accuracies during vehicle sessions. Administration of clinically relevant doses of amitriptyline significantly decreased task accuracy at the highest dose tested (3mg/kg). Administration of gabapentin increased mean task accuracy, though the effect was not statistically significant until intra-subject variability was reduced by selecting the individual best dose for each animal (which averaged 12.8mg/kg). Most of the effect was obtained during the presentation of long delay trials (18.2% above vehicle). Task improvement was sustained during sessions run 24h after gabapentin administration. In a series that used a task-relevant distractor to determine gabapentin's effect on attention, drug treatment reversed distractor-impaired accuracy during long delay trials (25.4% above vehicle). The selective improvement in long delay accuracy in both paradigms suggests improvement in encoding or retention components of working memory. It is currently unclear whether the ability of acute administration of gabapentin to modestly improve working memory occurs by a mechanism that could be related to its anti-allodynic mechanism of action.