This study aimed to analyze the effect of conventional, Problem-Based Learning (PBL), Numbered-Head Together (NHT), and integrated PBL-NHT learning models on the low- and high-ability students' critical thinking. This quasi-experimental study used a pretest-post-test nonequivalent control group design with a four-by-two factorial pattern. The analysis results showed that learners' critical thinking was impacted by learning models implemented in the classroom and influenced by the interaction between the learning models and learners' academic abilities. There was no difference in critical thinking among low-ability students who were taught with PBL, conventional, and integrated PBL-NHT, but a difference was found in students taught using NHT. There was no difference in critical thinking between high-ability students taught with conventional and NHT instruction, but there was a difference between students taught with PBL and integrated PBL-NHT instruction. To foster critical thinking, pupils with low academic ability should be taught using NHT, whereas those with high academic ability should be taught using PBL or an integrated PBL-NHT approach.