The prediction of pest regulation by multi-predator communities often remains challenging because of variable and opposite effects of niche complementarity and predator interference. Carabid communities are regulating weeds in arable fields and include a mix of species ranging from granivores to predators that are obligate omnivores. It is not clear from field studies whether granivore and obligate omnivore species either contribute equally or are complementary in the process of weed suppression, and little is known about the impact of potential predator interference within carabid communities on weed suppression. We compared the weed seed foraging strategy of the granivore Harpalus affinis and the obligate omnivore Poecilus cupreus. Using no-choice test experiments, we compared their activity and seed acceptance for four weed species through a scoring of the proportion of tested individuals consuming weeds, their latency before the consumption of the first seed and the total number of seeds consumed. We then evaluated their seed acceptance for dandelion seed Taraxacum officinale under predator interference by using chemical cues of carabids and tested the impact of three treatments, namely cues of intraspecific competition, interspecific competition and intraguild predation. We found that the obligate omnivore P. cupreus was highly active, had a low latency before consuming its first seed but had an interest in only two of the four weed species. P. cupreus seed acceptance remained unchanged in the presence of predator cues. By contrast, H. affinis was slow to start its seed consumption, accepted equally seeds of the four weed species and significantly increased its seed consumption in the presence of cues mimicking intraguild predation. These findings indicate that the two species differ in their foraging strategies, and as such, could have different contributions to weed seed suppression. This novel result calls for further studies documenting the foraging strategy of carabid species that thrive in arable fields as this could significantly improve our understanding of the delivery of weed seed regulation.