Testing boosts long-term memory relative to restudying (Bjork, 1975). Although using flashcards is regarded as a potent form of self-testing, little is known about conditions that affect flashcards’ effectiveness. One factor may be whether one uses flashcards in pairs or alone. Participants were randomly assigned to learn difficult vocabulary individually or in pairs. In the paired condition, one participant tested the other for 20 min, after which they reversed roles. In the individual condition, participants studied alone for 20 min and then tested themselves (or vice versa). Everyone then completed a final cued recall test. Though participants in the paired condition completed fewer test/study cycles, their final test recall equaled that of the individual condition. Further, paired flashcard use yielded more accurate predictions of final test performance than the individual condition, which overestimated performance by two letter grades. Thus, paired flashcard use is more efficient and improves metacognitive judgments. These findings may be especially crucial for students preparing for high-stakes exams, for which erroneous study decisions can have devastating consequences.