The main aim of the study, conducted in Poland, was to compare the academic achievement of 1552 (at Time 1) students without disabilities in three educational settings: general, without students with disabilities, inclusive (with co-teaching), with three to five students with disabilities, and inclusive (without co-teaching), with one to two students with disabilities. The study was longitudinal, with three waves eight months apart. The latent growth curve model was used for data analysis. The results have shown that changes in academic achievement (for language and for mathematics) over time were similar in all three groups; therefore, there were no differences between inclusive education (of the two types) and general education classrooms. These results mean that students neither lose nor benefit while learning in inclusive education classrooms. The findings are in line with the results of previous meta-analyses and are important for the future development of inclusive education.