<![CDATA[Mathematical problem-solving ability is a core competency in mathematics learning, yet many students still struggle to solve problems that differ from the examples given by the teacher because instruction remains dominated by conventional methods. This study is motivated by the low mathematical problem-solving ability of Grade VIII.3 students at SMP Negeri 12 Padang. The REACT learning model is considered a suitable alternative because its five stages connect mathematical concepts to real-life contexts and support each problem-solving indicator. The study aims to determine whether the mathematical problem-solving ability of Grade VIII.3 students at SMP Negeri 12 Padang after applying the REACT learning model is better than before its application, and whether there is an increase in students' mathematical problem-solving ability through the implementation of the REACT learning model. This is a pre-experimental research using a one-group pretest-posttest design. The sampling technique used was purposive sampling, with Grade VIII.3 students as the research subjects. The instrument used consisted of essay-format pre-test and post-test questions. The data analysis techniques employed were the t-test and N-Gain to measure the improvement. The results showed that the average pre-test score was 30.92, while the average post-test score was 57.52. The hypothesis test results at a significance level of a=0.05 obtained tcount=6.496 and ttable=2.040. Because tcount > ttable, the proposed hypothesis is accepted. Thus, the students' mathematical problem-solving ability after implementing the REACT learning model is significantly better than before its implementation. The gain score results showed an average increase of 26.61, categorized as moderate, while the N-Gain results showed an average score of 0.40, also falling into the moderate category. Therefore, there is a moderate improvement in students' mathematical problem-solving ability through the application of the REACT learning model.]]>
