INTRODUCTION: The correct diagnosis of autistic individuals is an everyday challenge within autism outpatient services. While the short-story task (SST), a task measuring fiction-based mentalizing, has demonstrated promise in differentiating between autistic and non-autistic adults, its discriminative ability has not been investigated in a sample of individuals seeking autism diagnostics at outpatient services. METHODS: This study aimed to evaluate the utility of the SST in individuals seeking autism diagnostics between 2016 and 2022 at the Clinic and Polyclinic for Psychiatry & Psychotherapy of the University of Regensburg at medbo District Hospital Regensburg. The sample consisted of 211 individuals. In 100 of them an autism spectrum disorder has been diagnosed and 111 individuals were diagnosed with other conditions or none at all. RESULTS: Performance on the SST did not significantly differ between the two groups, and receiver operator curve analysis did not support the SST as a reliable discriminator. However, linear regression analyses revealed that autism diagnosis was the sole significant predictor of SST mentalizing performance. Additionally, specific items of the SST showed significant differences between autistic and non-autistic individuals and constituted a significant predictor of autism diagnosis. DISCUSSION: While the SST may not be robust enough to accurately identify autistic individuals on its own, it does offer clinicians valuable insights into how individuals interpret others' actions and whether they grasp the broader context of a story versus focusing solely on details.