There are currently no FDA-approved medications to reduce cocaine relapse. The majority of preclinical studies aimed at identifying the neurobiology underlying relapse involve the self-administration of cocaine alone, whereas many, if not a majority, of cocaine users engage in polysubstance use. Here we developed a rat model of sequential cocaine and alcohol self-administration to test the hypothesis that this combination produces distinct neuroadaptations relative to those produced by cocaine alone. Male rats underwent intravenous cocaine self-administration (2 h/day) followed by 6 h access to unsweetened alcohol (20% v/v) for 12 days. After extinction training, we assessed surface expression of the glutamate transporter GLT-1 and glutamate efflux in the nucleus accumbens (NA) core during the reinstatement of cocaine-seeking. We also tested the ability of ceftriaxone to attenuate the reinstatement of cocaine-seeking and assessed reinstatement-induced Fos expression in several regions critical for reinstatement. Alcohol consumption did not alter cocaine intake, nor did access to cocaine alter alcohol consumption. However, we noted significant changes in glutamate homeostasis in the NA core of cocaine + alcohol rats relative to rats consuming cocaine alone, such as increased surface GLT-1 expression and a lack of increase in glutamate efflux during reinstatement of cocaine-seeking. A history of cocaine + alcohol also altered patterns of reinstatement-induced Fos expression. These changes likely account for the inability of ceftriaxone to attenuate cocaine relapse in cocaine + alcohol rats, while it does so in rats consuming only cocaine. As such glutamate neuroadaptations are targeted by medications to reduce cocaine relapse, preclinical models should consider polysubstance use.