There is a decrease in mathematics learning achievement and teachers have difficulty in determining appropriate learning models and strategies in the online learning process. Therefore, this study aims to identify the comparison of mathematical abilities based on an interactive learning model setting cooperative, expository, and problem-based learning. This research is a quantitative research with a quasi-experimental type conducted in one of the high schools in Jakarta. The inferential analysis technique is divided into two, namely the normality test using the Chi-Square test and the homogeneity test using the Levene test in the SPSS calculation program. The hypothesis test used by the researcher used the one-way ANOVA test because it used 3 variables and the data used were normally distributed and homogeneous. Furthermore, the researchers continued the paired hypothesis test t test. The results of the t-test resulted in a cooperative setting interactive learning model with expository learning and problem-based learning models having a significant difference in average mathematical ability. Cooperative setting interactive learning with problem-based learning model has a significant difference in average mathematical ability. Meanwhile, expository and problem-based learning model does not have a significant difference in average mathematical ability.