The teaching of physics, specifically the topic of heat, is essentially aimed at facilitating a deep understanding of concepts and enhancing pupils’ higher-order cognitive skills. However, the results of a preliminary field study indicate that the learning process is still hampered by the presentation of abstract concepts, meaning that pupils tend to simply memorise formulas without meaningful understanding. This study aims to analyze students’ cognitive abilities through the application of a guided inquiry learning model with a deep learning approach to the topic of heat. The research method used was a quantitative method with a between-group design, specifically a quasi-experimental research design. The specific research design employed was a pretest-posttest design with a non-equivalent control group. The research instrument used was a cognitive ability test consisting of thirteen items. Data analysis utilized the N-Gain score and the Mann-Whitney U test, which met the prerequisites of normality and homogeneity. The results showed that the N-Gain scores in the experimental class were higher than those in the control class, with values of 0.67 and 0.57, respectively. These results were supported by the Mann-Whitney U test, which yielded an Asymp. Sig. (2-tailed) value of 0.029 under the assumption of unequal variances (0.029 < 0.05). This indicates a significant difference in the improvement of students’ cognitive abilities between the experimental and control classes. Therefore, the guided inquiry learning model with a deep learning approach can serve as an alternative solution for enhancing students’ cognitive abilities.