This study investigates the interaction of task and personal characteristics during working with reality-based tasks. Real-world contexts with varying questions are created and administered to 535 9th and 10th-grade students in Germany. Generalized linear mixed models reveal that mathematical factors, especially the task contexts’ interestingness and familiarity, significantly impact solution rates. Moreover, socioeconomic status, mathematical self-efficacy, contextual prior knowledge, language competence, and intra-mathematical competence positively contribute to the rate of correct solutions. This research highlights the importance of considering not only mathematical aspects but also the task context.