Gifted underachievement represents a frustrating loss of potential for society. Although attempts have been made to develop interventions to reverse gifted underachievement, the theoretical underpinnings of these interventions have yet to be empirically validated. The purpose of this study was to investigate the validity of the Achievement-Orientation Model for gifted middle school students. Based on a sample of 156 gifted sixth and seventh grade mathematics students, results of the current study suggest two unique clusters of gifted students, those whose attitudes toward each of the constructs present in the Achievement-Orientation Model are positive and those who attitudes are negative. Significantly more gifted underachievers were found in the negative attitudes cluster and more gifted achievers in the positive attitudes cluster, χ2(1)= 15.86, p<.001. Further, only two of the constructs present in the model distinguished gifted achievers from gifted underachievers, self-efficacy, t(154)=-3.850, p<.001, d=.62 and self-regulation, t(154)=-3.113, p=.002, d=.50. Finally, results of a path analysis call into question several of the relationships specified by the model. In particular, task meaningfulness only predicted student engagement and was not significantly related to self-regulation or student achievement. The findings of this study suggest the Achievement-Orientation Model may hold promise for the development of interventions to address gifted underachievement; however, future research should be conducted to continue work toward validating the model before this step is taken.