Background: There is evidence indicating that instruction using ABRACADABRA (ABRA) – a free web application designed to promote literacy development – may benefit children with autism spectrum disorders (ASD) when administered on an individualized basis in children's homes. Aims: Here, we investigated the efficacy of ABRA instruction administered in small groups of children with ASD within a school setting. Methods and procedures: Children were aged 5.83–8.42 years (n = 23). Some children were assigned to an instruction group and received a minimum of 20 h of ABRA instruction over 9 weeks (n = 11). The other children comprised an age- and ability-matched control group (n = 12) and received business as usual literacy instruction. Outcome measures included word-level accuracy, passage-level accuracy, and passage-level comprehension, all assessed using standardized tests that were independent of ABRA. Outcomes and results: ANOVAs comparing pre- versus post-instruction raw scores showed statistically significant improvements in word- and passage-level reading accuracy for the instruction group relative to the control group, with large effect sizes. Gains in reading comprehension for the instruction group were not statistically significant and, in a posthoc correlational analysis, appeared to be related to children's socialisation skills (r =.62). Conclusions and implications: Literacy instruction using ABRA is associated with improvement in reading accuracy for children with ASD when administered in small groups within a school setting. Children with ASD may require additional supports to make gains in reading comprehension when literacy instruction using ABRA is delivered in groups.