This study was designed to determine whether opioid mixed agonist-antagonist analgesics other than buprenorphine also selectively reduce cocaine self-administration by rhesus monkeys. The effects of daily treatment with nalbuphine (0.1 to 3 mg/kg/day) or (0.254 to 7.62 mumol/kg/day), butorphanol (0.01 to 0.3 mg/kg/day) or (0.0209 to 0.628 mumol/kg/day), and saline on cocaine and food self-administration were each studied for 40 sessions over 10 consecutive days. Cocaine (0.05 or 0.10 mg/kg/inj) and food (1-gm banana pellets) self-administration were maintained on a fixed ratio 4 (variable ratio 16:S) schedule of reinforcement. Both nalbuphine and butorphanol reduced cocaine self-administration (p < 0.0001) but this effect was not selective since food self-administration also decreased in a dose-dependent manner (p < 0.0001). Nalbuphine administration (1 to 3 mg/kg/day) decreased cocaine injections to 40% to 60% below baseline (p < 0.01) and food pellets 30% to 68% below baseline (p < 0.01). Lower doses of nalbuphine (0.10 and 0.30 mg/kg) did not change cocaine- or food-maintained responding significantly. All doses of butorphanol (0.01 to 0.3 mg/kg/day) reduced cocaine injections to 16% to 58% below baseline (p < 0.01). Food self-administration decreased to 21% to 70% below baseline (p < 0.01) at butorphanol doses of 0.03 to 0.3 mg/kg/day). These data suggest that these opioid mixed agonist-antagonist analgesics may not be useful as pharmacotherapies for the treatment of cocaine abuse.