This study estimated ASD prevalence in a cohort of 3-year-old very preterm children (N = 55) and investigated the usefulness of parent-reported ASD screeners and the ADOS-2. 12.7% received an ASD diagnosis by clinical judgment based on DSM-5 criteria. An additional 14.5% were classified as having a broader-autism-phenotype outcome. Sensitivity values for the screeners were poor, whereas specificity values ranged from poor to excellent. The ADOS-2 identified all children with ASD and had a fair specificity. These findings confirm the elevated ASD prevalence made by previous studies with preterm children but also highlight the challenges of successfully identifying ASD in this at-risk group. Caution is warranted when interpreting results of ASD instruments with the currently available cut-off scores and algorithms, especially when developmental challenges are present.