The aim of this study is to assess whether a brief training with rhythmic beat gestures helps L2 pronunciation in a reading aloud task with high school students. In a between-subjects pretest-posttest design, a total of 59 high school students were randomly assigned to one of the following two conditions: the beat gesture group and no-beat gesture group. In the beat gesture condition they were asked to first read two short stories aloud without any gestural instruction (pretest) and in the following two texts they were asked to move their hands (training). Students in the no-beat condition (control condition) were asked to read the four texts aloud (pretest and training) without any gestural instruction. Then, in order to see the benefits of gesture, both groups were asked to read a fifth text aloud (posttest) which was more difficult (more complex syntactic structure and longer) than the ones they read in the pretest or the training. Results showed that speakers who were asked to produce beat gestures during the training had better pronunciation measures (specifically accentedness, comprehensibility, and fluency) in the posttest than the ones that were not asked to produce any specific gesture during the training.