Past work on media coverage of candidates for political office has explored gender differences in quantity, substance, and tone with mixed results depending on the office, race, and context. We draw on the stereotype content model (SCM) to examine gendered patterns of media coverage of candidates on the trait dimensions of warmth and competence in the 2020 U.S. Democratic presidential primary. Combining Natural Language Processing and manual analysis of news, we find that female candidates receive more negative than positive warmth coverage, while male candidates receive more tonally balanced warmth coverage, which suggests that female leaders are penalized on the dimension of warmth. Additionally, white women received more warmth coverage than women of color and women of color receive more competence coverage than white women. The findings suggest news media may portray white women and women of color candidates as lacking gender congruent traits like warmth but may portray white women as possessing role congruent traits like competence.