Four experiments compared the CNS effects of a novel M1/M3 receptor agonist L-689,660 with those of the M1/M3 muscarinic receptor agonist AF102B. In the mouse tail-flick test of antinociception (TF) the minimum effective doses to increase tail-flick latency (MED) of L-689,660 and AF102B were 0.03 mg/kg and 10.0 mg/kg, respectively. In a rat conditioned-suppression-of-drinking (CSD) test of reference memory, doses of 0.3 and 1.0 mg/kg L-689,660 and a dose of 5.0 mg/kg AF102B reversed a scopolamine-induced deficit in performance (0.6 mg/kg). Although there was a tendency for L-689,660 to reverse the scopolamine-induced (0.4 mg/kg) performance deficit in a rat delayed-matching-to-position (DMTP) test, the difference failed to reach statistical significance. In contrast, a 5.0 mg/kg dose of AF102B potentiated the scopolamine-induced deficit in choice accuracy and the number of trials completed on this task. In a response sensitivity (RS) test, chain-pulling rates were significantly decreased by L-689,660 (MED = 0.03 mg/kg) and by AF102B (MED = 5.0 mg/kg). These results suggest that L-689,660 and AF102B may ameliorate or reverse a scopolamine-induced deficit, but only at doses that also reduce chain-pulling rates on operant schedules of reinforcement.