Question 1

What actually differs between members of this class?

Accepted Answer

Four things, and a milligram figure is none of them. The first is which receptors a molecule occupies. A 2022 review of newer diabetes medicines describes tirzepatide as a combination of GIP- and GLP-1 receptor agonists, whereas most of the class acts on the GLP-1 receptor alone. The second is what the molecule is constructed from. A 2025 phase 3 report calls orforglipron a small-molecule, nonpeptide oral glucagon-like peptide-1 (GLP-1) receptor agonist, which is a manufactured chemical rather than a chain of amino acids. The third is approval status, and it moves over time. A 2025 review of real-world use restricts the approved weight-loss members of this class to liraglutide, semaglutide and tirzepatide, which leaves everything else investigational or unapproved. The fourth is the interval, and our dosage page for this class sets out which molecule goes in weekly and which goes in daily. A comparison skipping all four and setting one milligram figure against another compares labels rather than molecules.

Question 2

Is there a difference between GLP-1 and GLP-1RA?

Accepted Answer

Yes, and it is the reason this subject carries two names. A 2025 comparative review calls GIP and GLP-1 the two gut-derived incretin hormones, incretin meaning a hormone the intestine releases once food arrives. That is GLP-1 the hormone, which your own body already produces. GLP-1RA is shorthand for a GLP-1 receptor agonist, an engineered molecule built to occupy the same receptor. The shorthand covers more ground than it appears to. The same review notes that Semaglutide is a GLP-1 receptor agonist, while Tirzepatide acts as a dual agonist for the GLP-1 and GIP receptors. So a single abbreviation gathers molecules with different targets underneath it. GLP-1RA identifies which receptor is involved and says nothing whatever about which molecule is in front of a reader, which is why a comparison has to name the molecule before it can usefully name a number.

Question 3

Do GLP-2 and GLP-3 exist as medicines?

Accepted Answer

Not in this literature, and the numbering is a misreading worth clearing up. No source among the research behind this page describes a GLP-2 or a GLP-3 receptor agonist. The 1 in GLP-1 belongs to a hormone's name rather than marking a rung on a ladder of strength. What does exist beyond GLP-1 alone is a molecule adding a second or third target. A 2025 review describes emerging dual and triple co-agonists, targeting GLP-1 alongside glucose-dependent insulinotropic polypeptide, and glucagon pathways. A 2020 review sets out the engineering behind that, describing how GLP-1 activity can be built into the sequence of GIP. So the question has a real answer, although not the one its shape implies. There is no higher-numbered GLP anywhere in this record, and what a searcher is reaching for on typing GLP-2 or GLP-3 is a molecule carrying more receptors rather than a larger version of the same one.

Question 4

GIP and GLP-1: what does the second receptor add?

Accepted Answer

A different balance between two receptors, which a 2020 laboratory study set out to measure. It reports a greater degree of engagement of tirzepatide for the GIP receptor than the GLP-1 receptor, corroborating an imbalanced mechanism of action. Imbalanced here means the molecule leans harder on one of the two. Why anybody would want that arrangement is argued in a review from the same year. It points to findings in type 2 diabetes where dual GIP/GLP-1 receptor agonist therapy produces profound weight loss, glycemic control, and lipid lowering. Glycemic means relating to blood sugar, and lipid lowering means bringing blood fats down. Take the laboratory picture for precisely what it is. It is a reason to expect a difference rather than a measurement of one, and no study among the research behind this page traces a trial result back to receptor occupancy in people. The trials counted outcomes and the bench work counted receptors, and joining the two together is an argument rather than a finding.

Question 5

Daily tablet or weekly injection: what differs besides the route?

Accepted Answer

Three things: the molecule, the interval and the trial record standing behind each. The tablet side holds two quite different kinds of molecule. A 2019 phase 3a trial assigned adults with type 2 diabetes to once-daily oral semaglutide 3 mg, 7 mg, 14 mg, or placebo, which is a peptide rebuilt for swallowing. The other tablet is not a peptide at all. A 2025 phase 3 trial gave once-daily orforglipron at doses of 6 mg, 12 mg, or 36 mg to adults living with obesity. The weekly injection runs on a different clock again: a 2024 trial used a subcutaneous injection of either semaglutide 2·4 mg or placebo once a week for 44 weeks. Here is what the record cannot tell a reader. No trial among the research behind this page gives a tablet to one arm and an injection to another, so every comparison between those routes travels through separate placebo trials rather than through one. Separate trials enrol different people under different entry rules, and bridging that gap is precisely what an indirect comparison attempts.

Question 6

Which members of this class have been put in one trial together?

Accepted Answer

Three pairings, and they constitute the only direct evidence this class holds. A 2021 phase 3 trial randomly assigned 1879 patients, in a 1:1:1:1 ratio, to receive tirzepatide at a dose of 5 mg, 10 mg, or 15 mg or semaglutide at a dose of 1 mg. A smaller mechanism trial repeated that same pairing alongside a placebo arm. In it, patients were randomly assigned (3:3:2) to subcutaneously receive either tirzepatide 15 mg, semaglutide 1 mg, or placebo once per week. The third is still running. A cardiovascular outcomes trial randomized 1:1 to once weekly subcutaneous injection of either tirzepatide up to 15 mg or dulaglutide 1.5 mg, and its investigators write that it will provide definitive evidence as to the CV safety and efficacy of tirzepatide as compared with dulaglutide. CV is short for cardiovascular. Everything else said about one of these molecules against another is assembled from separate trials rather than observed inside a single room.

Question 7

Does a comparison made in diabetes carry over to weight loss?

Accepted Answer

Not automatically, and this is the trap inside most comparisons of this class. The one direct trial between two of its molecules was conducted in people with type 2 diabetes, and the primary end point was the change in the glycated hemoglobin level from baseline to 40 weeks. Glycated haemoglobin is the three-month average of blood sugar. The weight comparison was assembled somewhere else entirely. A 2026 pooled analysis of three molecules records that they were compared in adults without type 2 diabetes, and it reached them by combining separate trials rather than by conducting one. So the two most quoted comparisons in this class answer different questions in different populations. A blood-sugar result measured against one comparator does not transfer to a weight question in people that trial never enrolled, and treating the two as a single comparison is the commonest error made with these numbers.

Question 8

Why do two indirect comparisons reach different answers?

Accepted Answer

Because what gets pooled decides what comes out. A 2026 Bayesian network meta-analysis states its own filter: following a stringent heterogeneity assessment, six of 42 randomised controlled trials identified in the SLR were included in the NMA. SLR is short for a systematic review of the literature. NMA is short for a network meta-analysis, which lines up medicines that were never in one trial. A 2024 network meta-analysis kept far more of them. It reports that a total of 28 trials with 23,622 participants (44.2% female) were included, drawn from a different question in a different group of people. Read those two numbers before reading either conclusion. One analysis kept six trials and the other kept twenty-eight, so the two are not readings of one body of evidence. They are readings of two different bodies. A clash between them is a clash about what counts as comparable, not about what the molecules do.

Question 9

What changes when these are compared in ordinary practice?

Accepted Answer

The size of the gap narrows, and the reason is adherence rather than chemistry. A 2025 narrative review examined how the approved members of this class are used in practice, how they perform against each other, and what goes wrong with them. Its central observation about comparison is a shrinkage. The review reports that weight reduction in clinical practice tends to be lower than in randomized controlled trials. It adds that outcomes approach those seen in trials when focusing on highly adherent patients. That is an observation about two populations rather than about two molecules. A trial figure covers participants who satisfied the entry criteria and continued taking whatever they were handed, while a practice figure covers everybody who was given a prescription. Setting one molecule's trial number beside another's practice number compares the settings rather than the compounds, and that mismatch is where most of the disagreement originates.

Question 10

Are two GLP-1 molecules ever given together?

Accepted Answer

Not in this literature. No trial among the research behind this page gives two GLP-1 receptor agonists to the same participant, so there is no interaction, no interval and no outcome to report in either direction. What the record holds instead is one molecule carrying more than one target, and pairings with something from outside the class entirely. A 2024 article describes cagrilintide as an amylin-analog, now being developed in combination with the GLP-1 agonist semaglutide to achieve sustained weight loss in persons with overweight and obesity. Amylin is a separate hormone, so that pairing adds a different lever rather than doubling an existing one. The distinction is worth keeping hold of. A dual agonist is one molecule engineered to reach two receptors, whereas a stack is two molecules taken together by somebody who decided to. The record behind this page covers the first and has not examined the second, which leaves the second unmeasured rather than cleared.

Question 11

What has this class been measured against outside itself?

Accepted Answer

Placebo mostly, and two other kinds of medicine. A 2022 meta-analysis compared once-weekly tirzepatide 5, 10 or 15 mg with placebo or other glucose-lowering drugs in adults with type 2 diabetes, among them basal insulin regimens. Basal insulin is the long-acting background kind, given to hold blood sugar steady between meals. The second comparison is a pairing rather than a contest. A 2017 review of patients with type 2 diabetes argues that the SGLT2 inhibitors (SGLTi) and glucagon-like-1 receptor agonists (GLP-1 RAs) effectively reduce HbA1c, but via very different mechanisms, making them an effective duet for combination therapy. SGLT2 inhibitors are a separate class of diabetes medicine, and HbA1c is the three-month blood-sugar average. Each of those is a different kind of comparison. Collapsing them into one league table is how most readings of this class go wrong. Against placebo the question is whether anything happens at all. Against another diabetes medicine the question is which mechanism suits a particular patient, and alongside an SGLT2 inhibitor it is not a contest but an addition.

Question 12

How does this class compare with the research peptides on this site?

Accepted Answer

On the quantity of human evidence these are not the same kind of object at all. A 2022 meta-analysis inside this class reports that seven trials (6609 participants) were included. A 2025 phase 3 trial of one tablet in the class records that a total of 3127 patients underwent randomization. Most research compounds covered elsewhere on this site carry nothing of that size behind them. Their record is usually animal work, small uncontrolled series and a handful of early human trials, and that difference in scale is the honest headline of any comparison between the two groups. The word peptide is doing two jobs here as well. Several members of this class are peptides and one of the newest is not, so peptide describes how a molecule was constructed rather than which group it belongs to. Keep what the scale does and does not settle: a large trial record establishes that an effect was measured, in whom and how often something went wrong, while a thin record elsewhere means the question has not been asked yet rather than answered in the negative.

Question 13

What would a fair comparison between two of these need?

Accepted Answer

Four conditions, and the published record rarely supplies all four at once. The first is a single trial holding both molecules. The second is the same population. The third is each molecule at a comparable point on its own ladder, and the fourth is an endpoint both were designed to move. A 2026 pooled analysis says the first is missing for the three most familiar members, recording that head-to-head comparisons of all three of these interventions are lacking. It still reached a conclusion, by pooling trials that were never conducted against one another. The third condition is the one readers skip. That same analysis compared semaglutide 2.4 mg, liraglutide 3 mg and tirzepatide 5, 10 and 15 mg, which are three separate ladders with three separate top rungs. Setting the top of one against the middle of another compares trial design rather than molecules, and that is the sentence most GLP-1 comparisons leave out.

Trial	Arms compared	Ratio
2021 phase 3 trial, 1879 patients	Tirzepatide 5, 10 or 15 mg; semaglutide 1 mg	1:1:1:1
Mechanism trial	Tirzepatide 15 mg; semaglutide 1 mg; placebo, once per week	3:3:2
Cardiovascular outcomes trial (still running)	Tirzepatide up to 15 mg; dulaglutide 1.5 mg, once weekly	1:1

GLP-1 comparisons and stacks