A multiple baseline design across subjects with a reversal was used to examine the effects of classwide peer tutoring relative to traditional reading instruction on reading skills and social interaction time for 3 high‐functioning students with autism and their typical peers in integrated, general education classrooms. Traditional reading instruction consisted largely of teacher‐led instruction with individual student participation and seat work. Classwide peer tutoring consisted of 25 to 30 min of well‐specified instruction in which tutor—learner pairs worked together on a classwide basis on reading fluency and comprehension skills. All students participated in 15‐ to 20‐min unstructured free‐time activities immediately following reading instruction. Results of reading assessments demonstrated that classwide peer tutoring increased reading fluency and correct responses to reading comprehension questions for students with autism and their peers. The procedure further increased the total duration of free‐time social interactions for students with autism and typical peers, with individual variation