Based on the Implicit-Affect-Primes-Effort (IAPE) model (Gendolla, 2012, 2015), the present experiment investigated the role of prime visibility as a moderator of fear and anger primes’ effect on cognitive performance. Previous research has revealed inconsistent effects. Participants worked on a d2 mental concentration task with integrated pictures of fearful vs. angry faces, which were presented either masked (25 ms) or clearly visible (775 ms). Cognitive performance was assessed in terms of response accuracy and reaction times. Prime visibility significantly moderated the affect primes’ effect on response accuracy: When the primes were visible, fear expressions resulted in significantly lower response accuracy than anger primes. The opposite pattern occurred when the affect primes were masked. Additionally, visible primes led to slower responses in general, suggesting controlled prime processing. The observed performance effects corroborate recent findings on physiological measures of resource mobilization in the context of the IAPE model. Highlights Participants were presented with masked (25 ms) vs. clearly visible (775 ms) primes of anger or fear during a mental concentration task. The visibility of the primes significantly moderated the effect of affect primes on response accuracy. When the primes were visible, fear primes resulted in significantly lower response accuracy than anger primes. The opposite pattern occurred when the affect primes were masked. The performance results corroborate recent physiological findings related to the IAPE model.