The primary cause of brain tumors is the atypical proliferation of human brain cells. It is among the illnesses that kill the most people, both adults and children. There are various forms both benign (non-cancerous) and malignant (cancerous) brain tumors. Eighty to ninety percent of central nervous system (CNS) initial malignancies are brain tumors. Both benign (non-cancerous) and malignant (cancerous) brain tumors. Eighty to ninety percent of central nervous system (CNS) initial malignancies are brain tumors. Approximately 11,700 people are diagnosed with brain tumors each year. There are four other categories of brain tumors: pituitary tumors, glioma tumors, malignant tumors, etc. Appropriate care, preparation, and precise diagnosis are needed to extend the patients’ lives. The most effective method for identifying brain cancers is magnetic resonance imaging (MRI). The radiologist reviews the large amount of generated picture data. Because brain tumors and their qualities are complicated, a physical examination could be prone to errors. Artificial intelligence (AI) and machine learning (ML)–based automatic categorization methods have continuously surpassed manual classification in terms of precision. Therefore, to identify and categorize brain cancers, we suggest a lightweight model based on Convolution Neural Networks (CNNs). Using the MRI pictures, our CNN model will determine the type of brain tumor—glioma, pituitary, meningioma, or absent—from the scans. Brain cancers are also detected and classified using the pre-trained Visual Geometry Group 16 (VGG16) and Residual Networks 50 (Resnet 50), Inception v3 models. Lastly, we have evaluated the performance of our suggested CNN model against the pre-trained VGG16, Resnet 50, and Inception v3 models.