There is growing support for the disaggregation of psychopathy into primary and secondary variants. This study examines whether variants of psychopathy can be identified in a subsample (n = 116) of juvenile offenders with high scores on the Youth Version of the Psychopathy Checklist (PCL:YV). Model-based cluster analysis of offenders’ scores on the PCL:YV and a measure of anxiety suggested a two-group solution. The derived clusters manifested expected differences across theoretically relevant constructs of abuse history, hostility, and psychiatric symptoms. Compared with low-anxious primary variants, high-anxious secondary variants manifested more institutional violence, greater psychosocial immaturity, and more instability in institutional violence over a 2-year period, but similar stability in PCL:YV scores.