In-vehicle information system (IVIS) use is prevalent among young adults. However, their interaction with IVIS needs to be better understood. Therefore, an on-road study aims to explore the effects of input modalities and secondary task types on young drivers' secondary task performance, driving performance, and visual glance behavior. A 2 × 4 within-subject design was undertaken. The independent variables are input modalities (auditory-speech and visual-manual) and secondary task types (calls, music, navigation, and radio). The dependent variables include secondary task performance (task completion time, number of errors, and SUS), driving performance (average speed, number of lane departure warnings, and NASA-TLX), and visual glance behavior (average glance duration, number of glances, total glance duration, and number of glances over 1.6 s). The statistical analysis result showed that the main effect of input modalities is significant, with more distraction during visual-manual than auditory-speech. The main impact of secondary task types was also substantial across most metrics, aside from average speed and average glance duration. Navigation and music were the most distracting, followed by calls, and radio came in last. The distracting effect of input modalities is relatively stable and generally not moderated by the secondary task types, except radio tasks. The findings practically benefit the driver-friendly human–machine interface design, preventing IVIS-related distraction.