High-risk human papillomavirus (hrHPV) testing is important in cervical cancer screening and management algorithms. Roche (Pleasanton, Calif.) cobas hrHPV testing is commonly performed on both ThinPrep (TP) and SurePath (SP) samples, but performance of these platforms has not been fully investigated in the literature.Roche hrHPV testing was performed on 47,885 (TP = 18,295; SP = 29,590) out of 130,648 consecutive Papanicolaou tests, over 16 months; 1895 of those had interpretable biopsies.The overall hrHPV detection rates were similar in TP (13.5%) and SP (13.1%). The hrHPV positive rate was higher in SP (8.5%) than TP (7.3%, P0.0001) in women with negative cytology; the difference in other cytologic diagnosis categories was insignificant. TP samples had significantly fewer negative cytology diagnoses (7.3% versus 8.5%, P0.0001), more low-grade abnormalities in cytology and biopsies, and higher colposcopy referral rate (4.8% versus 2.7%, P0.0001) than SP. There were no differences between TP and SP in detecting ≥HSIL by hrHPV testing, cytology or biopsy. SP samples had a significantly higher rate of HPV 16/18 but a lower rate of non-16/18 hrHPV genotypes than TP.Roche cobas hrHPV testing was similar in both TP and SP platforms. The significantly lower hrHPV detection rate in cytological negative TP samples is likely related to higher cytology reporting rates for indeterminate and low-grade diagnoses in TP than SP samples. Significant differences were also observed in hrHPV genotyping results between TP and SP. Clinical risk stratification based on hrHPV testing may need to take testing platforms into consideration.