Diabetes mellitus is a challenging chronic health condition at the forefront of attention of the healthcare system. Important estimates quantifying how diabetes prevalence varies by age are available for the general population, but these estimates are poorly characterized among autistic adults. Improved diabetes prevalence and likelihood estimates are needed to understand the impact of diabetes on this population. We analyzed IBM MarketScan health claims data from 2019 to 2020 to estimate diabetes prevalence among privately insured autistic (N = 54,500) and non-autistic adults (N = 109,000) and estimate autistic adults' likelihood of diabetes relative to non-autistic adults at 5-year age intervals (e.g. 18-22 to 23-27) when controlling for demographic factors. Diabetes prevalence was 5.15% for autistic adults and 3.07% for non-autistic adults. Autistic adults had a higher likelihood of diabetes than non-autistic adults at most 5-year age intervals, with the greatest difference at ages 23-37 (odds ratio = 2.67; 95% confidence interval = 2.33-3.06). Our findings suggest that autistic adults may be more likely than non-autistic adults to experience diabetes. Thus, development of diabetes support services and programs that accommodate autistic adults' individualized needs are important for future study to promote optimal diabetes outcomes for this population.