Studies to date on the effects of benzodiazepines on neuropsychologic function have yielded conflicting data with respect to the type, severity, and duration of deficits that may be induced by these agents. As part of a placebo-controlled trial of alprazolam-XR (extended release) administered in combination with cognitive-behavioral therapy in patients with panic disorder, a battery of tests was used to measure neuropsychologic function. Thirty-eight outpatients were randomly assigned to receive either alprazolam-XR or placebo. Dosages were titrated up so that the alprazolam group (N = 18) received a mean dose of 4 mg/day (reduced in two patients because of sedative side effects). Neuropsychologic function after 6 weeks of therapy at the target dosage was compared with baseline assessments in each group. Both groups showed a statistically significant improvement from baseline to repeated assessments on measures of attention, executive functioning, psychomotor speed, and visual memory (p < 0.001); these gains were attributed to a practice effect. No significant changes were noted in measures of learning, verbal memory, or reaction time, and neither group showed any deterioration from baseline to retesting in any aspect of neuropsychologic function. These findings call into question the assumption that long-term benzodiazepine therapy produces significant neuropsychologic deficit in patients with diagnosed anxiety disorders.