Semaglutide vs Tirzepatide: What the Literature Shows

Semaglutide and tirzepatide are frequently named in the same breath, and published summaries often compress the distinction into a single phrase — “single versus dual agonist.” That phrase is correct but it leaves out most of what the primary literature actually distinguishes: two different peptide backbones, two different acylation chemistries, two different receptor-occupancy profiles, and two different molar masses that matter at the bench. This post sets the two molecules side by side as the published record describes them.

What is the core structural difference between semaglutide and tirzepatide?

The two peptides are built on different parent hormones. Semaglutide was described by Lau and colleagues in a 2015 Journal of Medicinal Chemistry paper as a human GLP-1 analogue carrying two amino-acid substitutions relative to native GLP-1 — Aib at position 8 and Arg at position 34 — and derivatised at lysine 26 (source 1). A 2024 systematic review of the semaglutide pharmacokinetic literature characterises the resulting molecule as roughly 94 percent homologous to native human GLP-1, with the modifications concentrated at positions 8, 26 and 34 (source 6). The backbone is therefore the 31-residue GLP-1(7–37) sequence with a short list of engineered changes.

Tirzepatide was built the other way round. Willard and colleagues, writing in JCI Insight in 2020, describe the molecule as having been discovered by engineering GLP-1 activity into the GIP sequence (source 3), a design first reported by Coskun and colleagues in Molecular Metabolism in 2018 under the development code LY3298176 (source 2). The result is a 39-residue peptide whose parent hormone is glucose-dependent insulinotropic polypeptide rather than GLP-1.

Incretin

A gut-derived hormone released in response to nutrient intake that potentiates glucose-stimulated insulin secretion. GIP and GLP-1 are the two principal incretins.

Mono-agonist

A ligand engineered to activate a single receptor target. Semaglutide is described in the literature as a selective GLP-1 receptor agonist (source 4).

Dual agonist

A single molecule that activates two distinct receptors. Tirzepatide is characterised as a dual GIP and GLP-1 receptor agonist (source 2).

Acylation

Conjugation of a fatty-acid chain to a peptide, used in both molecules to promote reversible albumin binding and slow clearance (sources 1, 3).

Which receptors does each peptide engage?

Semaglutide engages one receptor. The 2021 New England Journal of Medicine report of the SURPASS-2 trial refers to it plainly as “a selective GLP-1 receptor agonist” (source 4), and the discovery paper frames the entire optimisation exercise around GLP-1 receptor potency alongside albumin affinity (source 1).

Tirzepatide engages two, and published pharmacology reports that it does not engage them equally. Willard and colleagues reported that tirzepatide shows affinity for the GIP receptor comparable to that of native GIP, while binding the GLP-1 receptor with approximately five-fold weaker affinity than native GLP-1 (source 3). The same study described the molecule as imbalanced — greater receptor occupancy at GIPR than at GLP-1R at clinically studied exposures — and as biased at GLP-1R, favouring cAMP generation over β-arrestin recruitment, with weaker receptor internalisation than native GLP-1 (source 3). No comparable bias or imbalance framing applies to semaglutide, which has only one target receptor to be imbalanced between.

How do their binding affinities compare in published assays?

Direct numerical comparison across papers is limited, because the two molecules were characterised in different assay systems a decade apart. Lau and colleagues reported a GLP-1 receptor affinity of 0.38 ± 0.06 nM for semaglutide, a roughly three-fold decrease relative to liraglutide, and noted that the loss in receptor affinity was traded deliberately for a gain in albumin affinity (source 1). Willard and colleagues reported that in their low-expression-density cell system, the concentration–response curve for semaglutide was substantially right-shifted in the presence of human serum albumin, and calculated a dissociation constant of 1.86 µM for the tirzepatide–albumin interaction (source 3). The practical reading is that in both molecules, measured potency in a cell assay is strongly dependent on whether albumin is present, which is why free-drug estimates rather than total concentrations are used when receptor occupancy is modelled (source 3).

How does each molecule achieve a long circulating half-life?

Both use the same general strategy — a fatty diacid that binds albumin reversibly — but with different chain lengths. Semaglutide carries a C18 diacid (source 3), attached at lysine 26 through a linker designed to deliver both albumin affinity and stability against enzymatic degradation (source 1). Tirzepatide carries a C20 unsaturated diacid acyl chain (source 3).

The reported outcomes differ. Jensen and colleagues studied radiolabelled semaglutide in humans and reported that intact semaglutide was the primary circulating component, accounting for 69 to 83 percent of semaglutide-related material, with urine the primary excretion route and metabolism proceeding through proteolytic cleavage of the backbone and sequential beta-oxidation of the fatty-acid sidechain (source 7). The 2024 systematic review tabulated a terminal half-life of roughly 145 to 168 hours across volunteer studies — approximately one week — and reported values of 221 and 243 hours in subjects with severe renal impairment and end-stage renal disease respectively (source 6). For tirzepatide, Schneck and Urva pooled 19 studies into a population pharmacokinetic model, described the data with a two-compartment model with first-order absorption and elimination, and reported a half-life of approximately five days (source 5).

How do the two molecules compare on the published specifications?

What did the head-to-head study compare?

One published trial placed the two molecules in the same protocol. Frías and colleagues reported SURPASS-2 in the New England Journal of Medicine in 2021: an open-label, 40-week phase 3 trial that randomly assigned 1,879 enrolled subjects in a 1:1:1:1 ratio to tirzepatide at 5 mg, 10 mg or 15 mg once weekly, or to semaglutide at 1 mg once weekly (source 4). Baseline mean glycated haemoglobin was 8.28 percent and mean age was 56.6 years (source 4).

The investigators reported estimated mean changes in glycated haemoglobin from baseline to week 40 of −2.01, −2.24 and −2.30 percentage points in the three tirzepatide groups and −1.86 percentage points in the semaglutide group, with estimated between-group differences of −0.15 (95 percent CI −0.28 to −0.03), −0.39 (95 percent CI −0.51 to −0.26) and −0.45 percentage points (95 percent CI −0.57 to −0.32) respectively (source 4). Change in body weight was recorded as a secondary endpoint (source 4). Gastrointestinal events were the most frequently reported adverse events in the earlier phase 1 work on the dual agonist, and were described as dose-dependent and mild to moderate in severity (source 2).

How can a laboratory distinguish the two peptides analytically?

The molar-mass gap is the most direct handle. PubChem records semaglutide at a molecular weight of 4114 g/mol with the formula C₁₈₇H₂₉₁N₄₅O₅₉ (source 8) and tirzepatide at 4813 g/mol with the formula C₂₂₅H₃₄₈N₄₈O₆₈ (source 9) — a difference of roughly 699 daltons, far outside the mass tolerance of any competent mass-spectrometry identity check. A laboratory holding a vial labelled as one of the two therefore has an unambiguous confirmatory test available: the observed deconvoluted mass either matches the stated identity or it does not.

Purity and identity are reported separately on a Certificate of Analysis, and the two questions are answered by different instruments — typically HPLC for purity and mass spectrometry for identity. Every Steadfast Research Group batch ships with a batch-matched Certificate of Analysis so that the stated identity of the material in hand can be checked against the analytical record for that specific lot rather than against a generic document. Because the two peptides have distinct molecular formulas, distinct backbone lengths and distinct acyl chains, a mismatch between label and mass spectrum is straightforward to detect rather than a matter of interpretation.

What does the comparison literature not settle?

Several things remain open in the published record. The relative contribution of GIP receptor engagement to the differences observed between the two molecules has not been isolated experimentally in the head-to-head setting; SURPASS-2 compared whole molecules, not receptor pathways (source 4). Willard and colleagues framed the imbalanced-and-biased profile as a hypothesis that may account for the differences reported in early-phase work, explicitly presenting it as a proposed explanation rather than a demonstrated mechanism (source 3). Cross-study affinity comparisons remain difficult because the two molecules were characterised in different assay systems, with different albumin conditions, at different points in time. And the semaglutide pharmacokinetic literature itself reports a spread of terminal half-life values across populations and study designs rather than a single number (source 6), which is worth keeping in view when any single figure is quoted as definitive.