<![CDATA[The aims of this research are to analyze the improvement of understanding of mole concepts and chemistry calculations of class X students through the implementation of Chemistry Agent learning media. This study applies a quantitative approach through quasi-experimental methods and nonequivalent control group design. The research subjects include students in class X-2 as an experimental group and class X-4 as a control group at SMA Negeri 11 Surabaya in the 2025/2026 school year. Data was collected using a three-tier diagnostic test instrument in the form of pretest and posttest, which classified students conceptual understanding into three categories: scientific understanding, misconception, and lack of understanding. Data were analyzed by N-gain test, Shapiro-Wilk normality test, and Mann-Whitney test using IBM SPSS application. The results showed that the improvement in understanding of mole concepts and chemical calculations in the experimental class was higher than that of the control class with an N-gain value of the experimental class reaching 0.69 which was included in the medium category, while the control class obtained 0.47 also in the medium category. The results of the normality test showed that the data was not normally distributed, so the Mann-Whitney test was carried out and an Asymp value was obtained. A sig (2-tailed) of 0.005 (<0.05) indicates a significant difference between the experimental class and the control class. Furthermore,the three-tier classification showed that the experimental class achived higher percentages of conceptual understanding than the control class on almost all postest items, with conceptual understanding ranging from 45.16% to 83.87% in the experimental class and from 14.29% to 77.14% in the control class. Thus, the implementation of Chemistry Agent learning media is effective to improving the understanding of mole concepts and chemistry calculations of class X students of SMA Negeri 11 Surabaya.]]>
