Your VAR Model is Secretly an Efficient and Explainable Generative Classifier

Generative classifiers, which leverage conditional generative models for classification, have recently demonstrated desirable properties such as robustness to distribution shifts. However, recent progress in this area has been largely driven by diffusion-based models, whose substantial computational cost severely limits scalability. This exclusive focus on diffusion-based methods has also constrained our understanding of generative classifiers. In this work, we propose a novel generative classifier built on recent advances in visual autoregressive (V AR) modeling, which offers a new perspective for studying generative classifiers. Moreover, we show that the V ARbased method exhibits fundamentally different properties from diffusion-based methods. In particular, due to its tractable likelihood, the V AR-based classifier enables visual explainability via token-wise mutual information and demonstrates inherent resistance to catastrophic forgetting in class-incremental learning tasks. Generative models are trained to directly capture the underlying data distribution of a given dataset, which enables a wide range of applications such as image generation (Han et al., 2025), image editing (Mu et al., 2025), and data augmentation (Trabucco et al., 2023). Given this expressive capability, a natural question arises: Can we leverage these powerful generative models for classification? This question has motivated a line of research on the "Generative Classifier."