Our current knowledge of the implications of endocannabinoids in fear and anxiety is largely based on fear conditioning paradigms and approach-avoidance conflicts. Here we establish the ethobehavioral beetle mania task (BMT), which confronts mice with an erratically moving robo-beetle. With the help of this task we demonstrate decreased tolerance yet increased avoidance responses to an approaching beetle in high-anxiety behavior (HAB) and BALBc mice compared to C57BL/6N, CD1 and normal-anxiety behavior (NAB) mice. Also DBA/2N mice showed decreased passive and increased active behavior, but followed the robo-beetle more often than HAB and BALBc mice. Treatment with diazepam (1 mg/kg) increased tolerance without affecting avoidance behavior in HAB mice. Treatment with the MAGL inhibitor JZL184 (8 mg/kg) increased flight behavior, but did not affect tolerance. The FAAH inhibitor URB597 (0.3 mg/kg), however, reduced flight behavior and enhanced tolerance to the robo-beetle. The latter effects were blocked by co-treatment with the CB1 receptor antagonist SR141716A (3 mg/kg), which failed to affect the behavior by itself. Taken together, we validate the BMT as a novel test for studying endocannabinoids beyond traditional paradigms and for assessing active fear responses in mice. Furthermore, we demonstrate panicolytic consequences of pharmacological enhancement of anandamide, but not 2-AG signaling.