<![CDATA[This research is motivated by observations and interviews with teachers at MI NW Karang Bata, which showed that most students experienced difficulties in understanding story problems, determining solution steps, and selecting the correct arithmetic operations. Furthermore, students tended to be less active in the learning process and still required teacher guidance. Therefore, a learning model capable of actively engaging students in learning activities is needed, one of which is the Auditory Intellectually Repetition (AIR) learning model. This study aimed to determine the effect of using the Auditory Intellectually Repetition (AIR) learning model on the mathematical problem-solving abilities of fourth-grade students at MI NW Karang Bata in the 2025/2026 academic year. This study used a quantitative approach with a quasi-experimental design and a non-equivalent control group design. The study population was 39 fourth-grade students at MI NW Karang Bata. The sampling technique used saturated sampling, so the entire population served as the research sample. The research sample consisted of 21 students from class IV A as the experimental class and 18 students from class IV B as the control class. Data collection techniques were conducted through tests and documentation. The research instruments consisted of pre-test and post-test essay questions that had been tested for validity, reliability, difficulty level, and discriminatory power. Data analysis was conducted using descriptive statistics, followed by normality tests, homogeneity tests, and hypothesis testing using an independent sample t-test using the JASP application. The results showed that the Auditory Intellectually Repetition (AIR) learning model had a significant effect on students' mathematical problem-solving abilities. This was evidenced by the results of the hypothesis test, which obtained a p-value of 0.001 < 0.05, thus rejecting H₀ and accepting Ha₀. Furthermore, the average score of the experimental class was higher than that of the control class, at 79.29, compared to 64.44 for the control class. Thus, the AIR learning model is effective in improving students' mathematical problem-solving abilities.]]>
