kappa-Opioid agonists may functionally antagonize some behavioral effects of cocaine, but the role of mixed kappa/mu receptor activity is unclear. The effects of three mixed kappa/mu agonists (MCL-101, (-)cyclorphan, and Mr2034) and one kappa-selective agonist (enadoline) on cocaine self-administration and cocaine discrimination were compared in rhesus monkeys. Acute treatment with all kappa agonists dose dependently reduced cocaine-maintained responding and produced a downward shift in the cocaine self-administration dose-effect curve (0.001-0.32 mg/kg/inj, i.v.). During 7 days of chronic treatment, (-)cyclorphan (0.0032-0.032 mg/kg/h) and MCL-101 (0.0032-0.032 mg/kg/h) each dose dependently reduced cocaine self-administration maintained by a dose near the peak of the cocaine self-administration dose-effect curve. MCL-101 (0.032 mg/kg/h) produced selective and sustained decreases in cocaine self-administration, whereas (-)cyclorphan (0.032 mg/kg/h) had selective but transient effects. In addition, these mixed kappa/mu agonists produced fewer side effects (some salivation) than the kappa-selective agonist (sedation, salivation, emesis). However, none of these kappa agonists substituted for or antagonized cocaine's discriminative stimulus effects in monkeys trained to discriminate cocaine (0.4 mg/kg, i.m.) from saline. Thus, kappa and mixed kappa/mu-opioid agonists may reduce cocaine self-administration without altering cocaine's discriminative stimulus effects. Mixed kappa/mu agonists appear to offer some advantages over selective kappa agonists as potential treatments for cocaine abuse.