XX and XY fetal gonads are initially bipotential, poised between the ovary and testis fate. Multiple lines of evidence suggest that commitment to testis fate requires the repression of genes associated with ovary fate. It was previously shown that loss of CBX2, the subunit of the Polycomb Repressive Complex 1 (PRC1) that binds H3K27me3 and mediates silencing, leads to ovary development in XY mice and humans. While it had been proposed that CBX2 is an activator of the testis-determining gene Sry, we investigated the alternative possibility that CBX2 has a direct role as a repressor of the antagonistic ovary-promoting pathway. To investigate this possibility, we developed a quantitative genome-wide profile of the repressive histone mark H3K27me3 and its active counterpart H3K4me3 in isolated XY and XX gonadal supporting cells before and after sex determination. We show that testis and ovary sex-determining (SD) genes are bivalent before sex determination, providing insight into how the bipotential state of the gonad is established at the epigenetic level. After sex determination, many SD genes of the alternate pathway remain bivalent, possibly contributing to the ability of these cells to transdifferentiate even in adults. The finding that many genes in the Wnt signaling pathway were targeted for H3K27me3-mediated repression in Sertoli cells led us to test whether deletion of Wnt4 could rescue testis development in Cbx2 mutants. We show that Sry expression and testis development were rescued in XY Cbx2-/-;Wnt4-/- mice. Furthermore, we show that CBX2 directly binds the downstream Wnt signaler Lef1, an ovary-promoting gene that remains bivalent in Sertoli cells. Our results suggest that stabilization of the testis fate requires CBX2-mediated repression of bivalent ovary-determining genes, which would otherwise block testis development.