This chapter presents MobileViTNet, a novel hybrid deep learning model that integrates MobileNetV2 and Vision Transformer (ViT) for prostate cancer grade assessment using high-resolution pathology images. The dataset poses two significant challenges: the large size of the images, which demands efficient processing, and the presence of imperfect labels due to the inherent complexities of pathology. MobileViTNet tackles these challenges by utilizing MobileNetV2 for computationally efficient local feature extraction and ViT for capturing global dependencies and contextual relationships within the images. The hybrid model outperforms existing architectures, such as EfficientNet-B1, EfficientNet-B7, ResNet50, and DenseNet121, achieving a superior balance between accuracy and loss values while demonstrating strong generalization to unseen data. MobileViTNet’s ability to handle noisy labels and large-scale image data underscores its potential in medical image analysis, paving the way for improved diagnostic accuracy and consistency in prostate cancer grading. These findings highlight the transformative impact of artificial intelligence in healthcare, enhancing pathologists’ decision-making processes and ultimately improving patient outcomes.