Amphipods are ideal indicators for biomonitoring and ecotoxicological studies of environmental contaminants because they are extensively distributed in aquatic environments, are easy to collect and have an important role in nutrient cycling. Marine amphipods (Allorchestes compressa) were exposed to two concentrations of copper and pyrene, and their mixtures, for 24 and 48h. These contaminants were chosen because they are often detected together in coastal environments in close proximity to urban areas. Subsequent changes in polar metabolites were assessed using Gas Chromatography Mass Spectrometry (GC-MS)-based untargeted metabolomics. Generally, limited metabolic changes were observed for copper and pyrene single exposures, while significant metabolic changes were observed following exposures to mixtures. For all the exposure groups, although there were no significant variations in metabolite abundance between low and high exposures after 24 hours, there were eight metabolites that were significantly different to controls following copper exposure but only two metabolites were significantly different following pyrene exposure and all changes had seemingly returned to control levels after 48 hours. There were 28 metabolites that had significantly lower abundance compared to the control in the mixtures exposures after 24 hours, five of which were still significantly different after 48 hours. Multiple types of metabolites were affected following exposure, including amino acids, Tricarboxylic acid (TCA) cycle intermediates, sugars, fatty acids, and hormones.The results inferred the complicated dynamics of the interactions between amphipod detoxification mechanisms and contaminant exposures. Further studies would be beneficial to evaluate the dynamics of detoxification and recovery in long term exposures before metabolomics approaches can be used in biomonitoring programs. This study highlights the sensitivity of metabolomics in assessing the impacts of low concentrations of chemicals compared to traditional ecotoxicological endpoints.