In today's "obesogenic" environment, food cues signaling the availability of high-calorie food can trigger strong expectancies of eating, food cravings, and instrumental behaviors aimed at obtaining food. Such cue-elicited appetitive responses, collectively termed food cue reactivity, may lead to overeating, unsuccessful dieting, and weight gain. This study compared extinction and counterconditioning in reducing food cue reactivity, including outcome expectancies, craving, liking, and outcome-specific Pavlovian-to-Instrumental Transfer (PIT). We used a three-phase PIT paradigm: In the first two phases, participants learned Pavlovian and instrumental associations between two conditioned stimuli (CSs+) and two rewarding outcomes, as well as between two instrumental responses and the same outcomes. An additional stimulus was never paired with food outcomes (CS-). Participants then underwent either extinction or counterconditioning for one CS+ while the other CS+ remained unchanged, or they received no additional learning (control). In the final test phase, instrumental responding was measured in the presence and absence of the Pavlovian stimuli. In all phases, we measured outcome expectancies, craving and liking of the Pavlovian stimuli. Both extinction and counterconditioning reduced cue-elicited outcome expectancies, but only counterconditioning significantly decreased CS+ liking. Neither procedure effectively reduced cue-elicited craving. While outcome-specific transfer persisted after extinction, it was disrupted by counterconditioning, which led to a general decrease in instrumental responding. These findings suggest that counterconditioning may more effectively target the affective value of conditioned stimuli and reduce food cue reactivity, providing insights into potential interventions for weight management.