Recently, there has been a great deal of excitement about new glucagon-like peptide 1 receptor (GLP-1R) agonists (e.g., semaglutide and tirzepatide) that have received FDA approval for type 2 diabetes and obesity. Although effective, these drugs come with side effects that limit their use. While research efforts continue to focus intensively on long-lasting, orally administered GLP-1R medications with fewer side effects, a major impediment to developing improved GLP-1R medications is that the mechanism by which an agonist activates GLP-1R to imitate signaling is not known. Here we present and validate the G protein (GP)-first mechanism for the GLP-1R supported by extensive atomistic simulations. We propose that GLP-1R is preactivated through the formation of a GLP-1R-GP precoupled complex at the cell membrane prior to ligand binding. Despite a transmembrane helix 6 (TM6)-bentout conformation characteristic of activated GLP-1R, this precoupled complex remains unactivated until an agonist binds to elicit signaling. Notably, this new hypothesis offers a unified and predictive model for the activities of a series of full and partial agonists, including the peptides ExP5, GLP-1(7-36), and GLP-1(9-36). Most surprisingly, our simulations reveal an N-terminus domain (NTD)-swing/agonist-insertion mechanism wherein the long extracellular NTD of GLP-1R tightly holds the C-terminal half of the peptide agonist and progressively shifts the N-terminal head of the peptide to facilitate insertion into the orthosteric pocket. Our findings provide novel mechanistic insights into the activation and function of class B GPCRs and should provide a realistic basis for structure-based ligand design.