The objective of the current study was to apply untargeted metabolomics profiling to determine potential hypothalamic metabolite signatures unique to the most and the least feed-efficient animals based on residual feed intake (RFI) classification. One hundred forty-nine Red Angus cattle were allocated to 3 groups according to herd origin. Animals were fed a finishing diet for 78 d to determine the RFI category for each. Within each contemporary group, the 2 most-efficient (n = 6; RFI coefficient = −2.69 ± 0.58 kg dry matter/d) and least-efficient animals (n = 6; RFI coefficient = 3.08 ± 0.55 kg dry matter/d) were selected. Hypothalamic tissue was collected immediately after slaughter for metabolomics using a high-resolution mass spectrometry-based untargeted approach. Metabolites were analyzed using Q-Exactive MS system after LC separation. Data analysis was performed using the MetaboAnalyst 4.0 program. Metabolites with ‘importance in projection (VIP)’ scores >1.0 and a 2-fold difference between groups were considered significantly different. The top 15 metabolites with highest VIP score were identified by molecular weight (mass error ppm <5) for comparison between groups. There were 47 distinct metabolite features identified and a clear discrimination between groups revealed through multivariate analysis (PLS-DA and OPLS-DA). Among the top 15 metabolites identified by the VIP analysis, succinic anhydride (VIP = 3.4), tris(butoxyethyl) phosphate (VIP = 2.9), and etioporphyrin III (VIP = 2.7) were greater in the most-efficient compared with the least-efficient animals. Because of the well-established physiologic role of hypothalamus on feed intake control, the data indicate that untargeted metabolomics profiling could be helpful for identifying RFI-specific biomarkers that may play a role in determining feed efficiency in cattle.