Immersive technology with augmented reality (AR) as a didactic support has gone global and enriched the learning process with various packages of advantages. Although there have been many meta-analyses to test the aggregate effect of AR on students' academic performance, few have considered the duration of treatment as a moderator variable and also the comparison of the effect of AR in mathematics learning with other subjects. This study was conducted to. This random effect meta-analysis study was conducted to test the effectiveness of the application of immersive AR technology in learning by considering the duration of treatment and subject matter as research features and specifically identifying data from the Scopus database. This objective was achieved by examining 73 independent comparisons (n = 2822) that met the requirements and were identified from the Scopus database. The results of the CMA software-assisted analysis showed that the integration of immersive augmented reality technology in learning had a moderate effect (g = 0.75, p < 0.005) compared to learning conditions without AR. These results also add empirical validity to the relationship between categorical variables and the size of the research effect, as needed to understand research in the context of the application of AR in mathematics learning in the future. By clarifying the impact of AR implementation in mathematics learning, this study contributes to teachers to improve teaching effectiveness, enrich interactive learning media, and arouse students' interest and understanding of mathematical concepts in a more concrete and visual way. These findings also provide new directions for teachers, lecturers, stakeholders, and professionals in their efforts to develop a didactic framework by considering the duration of treatment in future AR applications.