GLP-1, "GLP-2" and "GLP-3": what people actually mean, and what you can actually get
Last checked August 24, 2026. This field moves faster than almost any other in medicine — verify before deciding anything.
First, the naming — because it trips almost everyone up.
There is no "GLP-2" or "GLP-3" weight-loss drug. The newer medications are not later versions of GLP-1; they act on more receptors at once. A single agonist hits GLP-1. A dual agonist adds a second target. A triple adds a third.
GLP-2 does exist as a hormone, and there is a real GLP-2 drug — teduglutide — but it treats short bowel syndrome and has nothing to do with weight. GLP-3 is not a thing at all. If a seller offers you "GLP-3", they do not know what they are selling.
The three generations people are usually talking about
| Drug | You may know it as | What it acts on | Trial loss |
|---|---|---|---|
| SemaglutideApproved in Canada | Ozempic, Wegovy, Rybelsus · the original “GLP-1” | Single agonist — GLP-1 onlyApproved and stocked. Generics since April 2026. | ~14–15% |
| TirzepatideApproved in Canada | Mounjaro, Zepbound · usually what “GLP-2” means | Dual agonist — GLP-1 + GIPApproved and stocked. No generic. | ~20–21% |
| RetatrutideStill in trials | No brand name — it is not on sale · usually what “GLP-3” means | Triple agonist — GLP-1 + GIP + glucagonNOT approved anywhere. Phase 3. Lilly planned an FDA filing in Q4 2026. | ~28% |
The quoted names in the second column are what people say, not what these are. They are in the table because that is how most readers arrive, and finding yourself in a table beats being corrected before you have got your bearings.
Percentages are average weight loss from each drug's own trials at the highest studied dose. They are not head-to-head except where noted — different trials enrol different people for different lengths of time, and comparing across them overstates how precisely they can be ranked. The same caveat applies to every figure on this page.
What each receptor actually does
This is worth understanding properly, because it explains why the drugs differ in more than strength — they differ in what they do.
GLP-1 — the appetite and insulin one (semaglutide — the original "GLP-1")
GLP-1 is a hormone your gut releases when food arrives. A GLP-1 drug is a long-lasting copy of it, and it does four things at once:
- Tells the pancreas to release insulin only when blood glucose is actually high. That glucose-dependence is why these rarely cause hypoglycaemia, unlike older diabetes drugs that push insulin regardless.
- Suppresses glucagon, which would otherwise raise blood sugar.
- Slows stomach emptying, so a meal sits longer and fullness lasts — this is also where most of the nausea comes from.
- Acts directly on appetite centres in the brain. This is the part people describe as the "food noise" going quiet: not willpower, but the constant background pull toward eating switching off.
+ GIP — the fat-tissue and tolerability one (tirzepatide — what people call "GLP-2")
GIP is the other incretin, released from a different part of the gut. Adding it (tirzepatide) contributes more glucose-dependent insulin release, but the interesting effects are elsewhere: GIP acts on fat tissue itself, appearing to improve how efficiently it stores and buffers incoming fat, and with it insulin sensitivity.
It may also act in the brain to reduce nausea, which is the leading explanation for something otherwise odd: tirzepatide is both more effective than semaglutide and frequently better tolerated. In a direct trial, 20.2% against 13.7%.
An honest wrinkle worth knowing, because it shows how young this science is: drugs that block the GIP receptor also produce weight loss. MariTide pairs a GLP-1 agonist with a GIP antagonist and works. Both directions helping is not fully explained, and anyone telling you confidently why GIP works is ahead of the evidence.
+ glucagon — the burn-it-off one (retatrutide — what people call "GLP-3")
This is the counterintuitive addition. Glucagon is insulin's opposite — it raises blood sugar, which sounds like the last thing you would want in a diabetes or obesity drug. Two things make it worth including anyway:
- It increases energy expenditure. Every drug above works mainly by reducing intake. Glucagon adds the other side of the equation — burning more, not only eating less — which is most likely why retatrutide's numbers step up rather than inch up.
- It drives the liver to burn its own fat. This is the direct explanation for the liver results below, and it is a genuinely different mechanism from simply losing weight and the liver improving as a consequence.
The engineering trick is the pairing: glucagon pushes glucose up, GLP-1 pulls it down. Combined in the right ratio you keep the energy-expenditure and liver effects while the glycaemic penalty is cancelled out. That balance is the whole design, and it is also why the dosing of a triple agonist is harder than a single one — get the ratio wrong and the glucose control goes with it.
What each of those costs you
Side effects follow the mechanisms directly, which is why they get worse in roughly the order above. The gastrointestinal effects are all downstream of slowed stomach emptying — the same action that produces the fullness — so they are not a separate problem you might avoid; they are the treatment being felt.
| Drug | Most common | Heart rate | Worth knowing |
|---|---|---|---|
| Semaglutide | Nausea (~44%), diarrhoea (~30%), vomiting (~24%), constipation (~24%) | +3–5 bpm | The GI effects are the slowed gastric emptying being felt. Worst on the way up. |
| Tirzepatide | Nausea (~25–30%), diarrhoea (~23%), vomiting (~13%), constipation (~11%) | +2–5 bpm | Consistently lower GI burden despite being the stronger drug — the GIP-reduces-nausea hypothesis is the leading explanation. |
| Retatrutide | Nausea (~43% at 12 mg), plus the same diarrhoea/vomiting/constipation pattern | +5–10 bpm | The heart-rate rise is the largest of the three and is attributed to glucagon, which speeds the heart directly. Dysesthesia — odd skin sensations, tingling — was reported by ~21% at 12 mg and is not a feature of the other two. |
Rates are from each drug's own trials at the highest dose and are not comparable across columns for the same reasons the weight figures are not. Almost all of it is worst during dose increases and settles at a steady dose — which is the practical argument for titrating slowly, and the reason people who rush it conclude they cannot tolerate a drug they could have tolerated.
Rare but serious, and shared across the class: pancreatitis, gallbladder disease (more likely with rapid weight loss), and a thyroid C-cell tumour signal from rodent studies that has not been established in humans but keeps these contraindicated in anyone with a personal or family history of medullary thyroid carcinoma or MEN2.
The one that is a pill
Orforglipron is not more powerful — its trial weight loss sits below injectable semaglutide. What it changes is the barrier: a small-molecule tablet with no food or water timing restrictions, and no cold chain. The FDA approved it in April 2026 as Foundayo. Health Canada has it under review, with approval plausibly late 2026 or 2027. For people who will not inject, a drug they will actually take beats a stronger one they will not.
Retatrutide, and why the numbers are not the whole story
Retatrutide's Phase 3 results are the largest ever recorded for an obesity drug — around 28% average weight loss, with a substantial share of participants losing over 30%. It is also not approved anywhere in the world. Lilly planned to file with the FDA in Q4 2026; a 10–12 month review puts plausible US approval in late 2027 and launch in 2028, with Canada typically behind that.
The early results beyond weight — and they are genuinely early
The weight number gets the attention, but the trial programme is testing it across several conditions at once, and some of what has reported is more interesting than the scale reading:
- Fatty liver disease (MASLD). A Phase 2a trial published in Nature Medicine reported liver fat falling by around 82% at 24 weeks on the highest dose, with 86% of participants reaching a normal liver-fat level (under 5%) — and up to 86% reduction by 48 weeks. This is the glucagon arm doing what glucagon does. A dedicated Phase 3 is running.
- Knee osteoarthritis pain. TRIUMPH-4, the first Phase 3 to report, paired 28.7% weight loss with roughly 75% pain reduction at 68 weeks. How much of that is load coming off the joint versus anything else is not separable from this trial.
- Sleep apnoea, type 2 diabetes, cardiovascular outcomes. All under study within the same programme, with further readouts expected through 2026.
Hold all of this loosely. The liver result is Phase 2a — small, early, and the stage at which drugs most often look better than they turn out to be. Cardiovascular outcome trials, the ones that take years and actually count events, have not reported. Impressive early findings are how most promising drugs look right before the larger trials either confirm them or do not, and a good number do not.
Which means anything sold as retatrutide today is an unapproved drug from an unregulated supplier, with no verified purity or concentration and no recourse if a vial is wrong. The trial numbers were produced with pharmaceutical-grade material under medical supervision; they are not a claim about what is in a vial bought online. That gap is the entire risk, and it is not visible from the outside.
The lesser-known ones, which is where the 1-2-3 idea falls apart
None of these fit a numbered sequence. One is a GLP-1 in a tablet. One pairs a GLP-1 with a hormone system that is not on the list at all. One is a dual agonist with a different second target from tirzepatide's. They are not a later generation of anything — they are different combinations being tried at the same time.
| Drug | You may know it as | What it acts on | Trial loss |
|---|---|---|---|
| OrforglipronUnder review | Foundayo · the GLP-1 pill | Single agonist — GLP-1, but a tablet rather than an injectionNOT approved. Under Health Canada review; approval plausibly late 2026–2027. | ~11–12% |
| CagriSemaUnder review | No brand name yet | Semaglutide + cagrilintide — a GLP-1 plus an amylin analogue, which is a different hormone system againNOT approved. Filed with the FDA Dec 2025; a US decision was expected late 2026. | ~22–23% |
| SurvodutideStill in trials | No brand name yet | Dual agonist — GLP-1 + glucagon, a different pair from tirzepatide'sNOT approved. Phase 3 running; approval projected 2027–2028 at the earliest. | ~19% (Phase 2) |
What this means if you are choosing today
In Canada, right now, the real choice is between semaglutide (cheapest by a distance since generics landed) and tirzepatide (more effective, no generic, more expensive). Everything else on this page is a reason to revisit the decision in a year, not an option available this month. Waiting for a better drug has a real cost too — the years spent waiting are years untreated — and that tradeoff belongs to you and your prescriber, not to a table.
Prices and access for what you can actually get in Canada are in the cost and access guide. If you are already on something, Selfsense tracks your doses, sites and weight against every dose change.
General information, not medical advice, and not a recommendation of any treatment. Trial results describe averages in studied populations, not what will happen to you. Prescribing decisions belong to you and a licensed healthcare professional.