Tumor-associated macrophages (TAMs) are associated with poor prognosis in multiple solid tumor types. Targeting signaling pathways that re-program TAMs from an immunosuppressive “tumor-promoting” (“M2-like”) to a pro-inflammatory “tumor-fighting” (“M1-like”) phenotype have shown promise for treating cancer in pre-clinical animal models1. However, in order to find the best method to reprogram macrophages, we need a better understanding of the molecular pathways that control tumor-associated macrophages1. The BAF (Brg/Brahma associated factor) complex is a multi-subunit nucleosome-remodeling complex that controls cell type-specific chromatin accessibility and gene expression, including regulating inflammatory gene networks2. To study the role of the BAF complex in tumor-associated macrophages, we used a mouse model to delete ARID1A3 (the largest subunit of the BAF complex) in the myeloid lineage using LysM-Cre4. ARID1A deletion had no effect on TAM accumulation following subcutaneous injection of MC38 colon cancer cells, but did cause an enrichment of an “M1-like” inflammatory gene expression signature, including upregulation of Cd86 and Cd274, the gene encoding the immunosuppressive ligand PD-L15,6. ARID1A-deficient TAMs also exhibited elevated CD86, and PD-L1 expression by flow cytometry. We tested the effects of anti-PD-L1 blocking antibodies and found that tumors were smaller in mice with myeloid-specific deletion of ARID1A, indicating heightened sensitivity to anti-PD-L1 therapy compared to control mice. Increased sensitivity to PD-L1 was correlated with increased cells surface expression of MHC class II and CD86 on TAMs and increased infiltration of CD8 T cells in myeloid-specific ARID1A-deficient hosts, a smaller proportion of which were ‘terminally exhausted’ CD8 T cells. TAMs represent a significant barrier to successful cancer therapy, however, also provide a promising opportunity for intervention with novel therapeutics. Our results demonstrate that ARID1A is a critical regulator of TAM polarization and immunosuppression and suggest that ARID1A could be a potential target to increase response to existing anti-PD-L1 immunotherapies.