With the focus on quality assurance practices during pavement construction, the present research aims at investigating the compactability of hot mix asphalt using the ground-penetrating radar technique. Thus, density as an indicator of the compactability of hot mix asphalt is predicted using three different electromagnetic-mixing-theory-based density models (namely, the complex refractive index model, Rayleigh model, and Al-Qadi, Lahouar and Leng model), and the prediction performance is also investigated. The investigations are based on experimental data acquired, both in the laboratory and field, from new full-scale asphalt pavement sections with varying asphalt mixture compositions. The laboratory experiment, which involved the compaction of asphalt mixtures using the steel-segmented roller compactor, indicated that compaction mode affects the compactability of hot mix asphalt, whereas the analysis of field ground-penetrating radar experimental data revealed that the estimated electric permittivity ε HMA during the compaction process could be considered a measure of the asphalt mix field compactability. The prediction performance of the three density models was evaluated using different methodological approaches with respect to the backcalculation of ε s of the mix aggregates. The results indicated that, by utilizing the ground-penetrating radar field measured ε HMA for the assessment of ε s , the predicted G mb values from the implementation of the above density models closely approach the ground-truth field-core bulk densities. Comparative evaluation of the three density models showed that the Al-Qadi, Lahouar, and Leng model exhibits the best prediction performance, which is comparable to nuclear/non-nuclear methods. In light of this, it could be argued that the ground-penetrating radar methodology coupled with novel algorithms can be an effective and efficient tool to improve the asphalt mix compaction process and assessment of in situ density.