The 2013 curriculum requires students to learn from anyone including their peers. There are several learning models that emphasize this, including Jigsaw and TSTS (Two Stay Two Stray). This study aims to compare learning results from the application of the two models to students of class XI of SMAN 1 Ambarawa. Sampling with cluster random sampling resulted in a sample of students of class XI MIPA 1 as an experimental class and students of class XI MIPA 2 as a comparison class. A quasi-experimental study with the design of The Randomized Control Group Pretest-Posttest Design was carried out on mathematics learning on the material derivative from algebraic functions. Test different mean initial ability of students by using the Mann-Whitney test produces a significant value of 0.443 (more than 0.05), meaning that the initial conditions of both classes are balanced. The hypothesis test with the Mann-Whitney test produced a significant amount of 0.029 (less than 0.05), with the average in the experimental class (76.81) higher, than the comparison class average (73.79). Therefore, it can be concluded that the application of jigsaw produces the results of mathematics learning in class XI students of SMAN 1 Ambarawa better than the application of TSTS