The present study tested the hypothesis that there are two broad groups of problem drinkers: the Apollonian-Dionysian distinction. Apollonian drinkers are defined by a later onset of alcohol problems, a slower developmental course, and less problem severity. Dionysian drinkers are defined by more severe alcohol problems, an earlier onset, and worse prognosis. The discriminant validity and classification agreement of five operations of the Apollonian-Dionysian model were tested in a general population (n=8643) and community sample (n=664). A κ-means cluster analysis supported the Apollonian-Dionysian distinction. Dionysian problem drinkers were more likely to be males, report greater alcohol consumption, more alcohol related antisocial behavior, have a current DSM-IV alcohol dependence diagnosis, and comorbid depression. Apollonian problem drinkers were more prevalent in the general population and reported less severe drinking problems. Classification agreement was stronger for subtype models that incorporate dimensions of alcohol use and/or the frequency of negative consequences. Babor's Type A-Type B model demonstrated the strongest reliability and consistency over time.