The face inversion effect—where face recognition performance declines dramatically for inverted faces—has long been central to understanding face perception. While some studies argue that upright and inverted faces are processed qualitatively differently, with only upright faces processed holistically, others suggest the difference is quantitative, proposing that holistic processing occurs for faces in both orientations. A recent framework, the “two faces of holistic processing,” offers a promising resolution by identifying two distinct aspects of holistic processing: facilitation, when irrelevant facial parts aid target processing, and interference, when irrelevant parts hinder it. To examine whether differences between the upright and inverted face processing are qualitative or quantitative, we employed the complete composite face task, which simultaneously measures facilitation and interference. Across multiple experiments, participants (N=40 in Experiment 1a, N=40 in Experiment 1b, N=120 in the preregistered Experiment 2) viewed faces in both upright and inverted orientations. Composite effects were consistently observed for both orientations, with stronger effects for upright faces, suggesting a quantitative difference in holistic processing. Similar patterns emerged for facilitation and interference when sensitivity *d’* and response times were analyzed separately. Crucially, exploratory analyses using inverse efficiency scores, which integrate accuracy and response times to account for potentially varied speed-accuracy tradeoffs in individuals, revealed that interference was reliably observed only for upright faces, whereas facilitation was comparable across both orientations. This pattern indicates that interference differs qualitatively, while facilitation differs quantitatively or not at all between upright and inverted faces. Together, these findings help resolve a long-standing debate in the face perception literature by showing that upright and inverted faces diverge primarily in interference, while sharing similar mechanisms for facilitation. The results provide further support for the two faces framework of holistic processing and highlight the importance of distinguishing between facilitation and interference in holistic face perception.