Previous studies have found little or no systematic differences in the rates at which female and male scientists are awarded funding in international grant competitions. However, past investigations have only studied outcomes, not the preceding scoring and selection process. We propose that common grant review practices–such as panel deliberations, score binning, and interview assessments–allow unequal evaluations to be corrected while staying within a framework of merit-based review. We analyzed unique data from a large funding competition, the Netherlands’ Organization for Scientific Research's Talent Program, including reviewer and panel evaluation scores of both funded and unfunded proposals. We replicate prior research demonstrating gender equity in funding outcomes. At the same time, we find that men received higher evaluation scores, consistent with our argument. This gender difference is counteracted by panels funding women with lower scores than men's, redistributing 64 million euro back to women that would otherwise have gone to men. Our study thus reveals that female scientists are more poorly evaluated than their male counterparts in spite of what equality in outcome statistics might suggest.