Abstract <p>This study presents a comprehensive evaluation of four widely used constitutive models—Hollomon, Ludwik, Swift, and Voce—in predicting the flow behavior of AA2024 aluminum alloy under both tensile and compressive loading conditions. True stress–strain curves were constructed from experimental tensile test data and extended to high-strain conditions (true strain up to 3.0) to assess their predictive accuracy through numerical simulations of the Rastegaev compression test. At&#xa0;small strains, the Ludwik model demonstrated the highest accuracy (<i>R</i>² = 0.99414), followed by the Swift model, while the Voce and Hollomon models showed relatively lower accuracy. However, in the large-strain regime, the Voce model exhibited superior predictive performance, particularly due to its ability to represent stress saturation, with the lowest simulation error in the Rastegaev test. These findings highlight the importance of selecting appropriate constitutive models based on strain regime and loading type. The results provide valuable insights for improving the reliability of finite element simulations in bulk forming processes involving lightweight aluminum alloys. The validated models can support the optimization of industrial forming operations and enhance material behavior prediction in large-deformation conditions.</p>