Cervical cancer remains a significant worldwide health challenge, particularly in developing nations, where early diagnosis is crucial for effective treatment and improved survival rates. Histopathological analysis of cervical tissue through Whole Slide Images (WSIs) plays a pivotal role in detecting Cervical Intraepithelial Neoplasia (CIN). The manual diagnosis is time-consuming, subjective, and sensitive to inter-observer variability. It provide a hybrid deep learning model in this work that combines the advantages of vision and convolutional neural networks (CNNs), particularly ResNet to perform accurate cervical cancer prediction from patch to WSI level. Our approach integrates low-level feature extraction through ResNet with ViT’s global attention capabilities to capture complex spatial relationships across histological structures. The model gets trained and evaluated on available publically accessible datasets such as TCGA and PAIP, focusing on CIN classification and early detection. Extensive experimentation demonstrates that the hybrid model outperforms standalone CNN or ViT architectures are better when it comes to accuracy, AUC-ROC, and F1-score. This work presents a promising AI-assisted diagnostic framework for automated and reliable cervical cancer screening.