The gut-brain axis is a bi-directional communication system through which microbial communities in the gut interact with the nervous system. Disruptions in microbiome composition, known as dysbiosis, appear to be associated with neuropsychiatric disorders, perhaps including drug abuse. This study used behavioral data and biological samples from the Cocaine Biobank to test the hypothesis that the gut microbiota can predict and reflect susceptibility to cocaine reinforcement. Adult male heterogenous (HS) rats were catheterized and allowed to self-administer cocaine in daily short-access sessions (2 hr/day, 10 days, 0.5 mg/kg per intravenous infusion), followed by progressive ratio (PR) testing. Rats were transitioned to daily long-access sessions (6 hr/day, 14 days), followed by a PR test and alternating blocks of footshock testing, long-access, and PR. Fecal samples were collected at three time points and bacterial 16s rRNA genes were sequenced to profile the microbiota and compare low vs. high responders to cocaine. Bacterial taxa identified in baseline samples in drug-naive animals were used to test whether specific bacterial abundance could predict future cocaine susceptibility. As expected, levels of cocaine-related behavior varied across individual subjects, such that a quartile split identified low and high responders on each measure, as well as an overall addiction index. Although beta diversity in the gut microbiota at baseline and after short access did not predict membership in high or low addiction quartiles, linear discriminant analysis (LDA) identified certain taxa that were more robustly represented in either low or high responders. Beta diversity after long access did reveal a difference in microbiota profiles between low and high responders, and LDA identified specific populations that differed between groups. Plotting baseline samples identified using LDA on a Receiver Operating Characteristic (ROC) curve revealed that high relative abundance of Akkermansia muciniphila predicted future low response rates. This study is the first to report that microbiota variability reflects levels of cocaine intake and that the microbiota might facilitate diagnosis and identify risk factors predictive of future drug vulnerability.