Colorectal cancer (CRC) exhibits substantial heterogeneity in response to immunotherapy, highlighting the critical need for improved stratification strategies. Here, we established a robust multi-omics framework by integrating transcriptomic, epigenetic, and mutational profiles, identifying three cancer subtypes (CS1-CS3). Among these, the CS3 subtype emerged as a distinct immunogenic class characterized by genomic instability and robust immune activation, capturing a subset of immune-inflamed tumors classified as clinically microsatellite-stable (MSS) that may be overlooked by routine microsatellite instability (MSI) testing. Single-cell transcriptomics further supported this immune-active state by revealing enrichment of pro-inflammatory myeloid populations and specialized epithelial states. Longitudinal analysis of independent immunotherapy cohorts revealed that CS3 signature activity exhibits response-dependent dynamics. While elevated baseline CS3 activity was associated with favorable clinical outcomes, significant post-treatment score reductions were primarily enriched among patients achieving radiographic complete response (CR). To facilitate clinical translation, we derived an 11-gene prognostic model for risk stratification and leveraged pharmacogenomic screening to prioritize actionable combination candidates, including multi-kinase and epigenetic inhibitors. Collectively, these findings characterize a therapeutically relevant immune-active subtype and provide a comprehensive framework of biomarkers and actionable targets to advance precision immunotherapy in CRC.