Simultaneous interpreting is generally assumed to be particularly demanding with respect to cognitive resources such as attention and working memory, which are thought to gradually increase with professional practice. Experimental data to corroborate this assumption is still rather sparse, however. Here we report an in-depth investigation of working memory capacity among 21 professional interpreters (experts), 18 second-year interpreting students (novices) and two control groups (20 multilinguals and 20 students). Tests involved either short-term retention alone; short-term retention and processing in a recall task with articulatory suppression, a listening span task, and a category and rhyme probe task; or attention alone in a unilingual and bilingual Stroop test. No between-group differences in simple span tasks and the Stroop test were found. Significant group effects were observed in free recall with articulatory suppression, in the category probe task and in the listening span task. The best performance was always produced by the novice interpreters rather than by the experts. These findings are discussed in relation to (a) the novice–expert distinction and the role of working memory in the development of interpreting skills, and (b) the nature of the task and possible strategies involved.