Semaglutide vs Tirzepatide: What Two Head-to-Head Trials Actually Settled

Last reviewed 28 August 2026 · Educational, not medical advice · No sponsorships, no affiliate links

What actually separates the two molecules

Semaglutide activates one receptor: GLP-1. Tirzepatide activates two: GLP-1 and GIP. Both are peptides, both are injected weekly, both slow gastric emptying and blunt appetite through overlapping central pathways.

The second receptor is the whole argument. GIP agonism appears to add weight loss beyond what GLP-1 alone achieves, and - counter to what most people assume about a drug that does more - it does not appear to add proportionally more nausea. That combination is unusual in pharmacology, and it is the reason the head-to-head result went the way it did rather than trading efficacy against tolerability.

Everything else people argue about is a brand or dosing question, not a molecule question. Ozempic and Wegovy are the same semaglutide at different ceilings; Mounjaro and Zepbound are the same tirzepatide at identical ceilings. We cover those separately in Ozempic vs Wegovy and Wegovy vs Zepbound.

The only fair comparison: two head-to-head trials

Almost every comparison you will read lines up numbers from separate trials with different populations, baselines, durations, and statistical conventions. That arithmetic is not valid. Two trials put the molecules against each other directly, and only those two settle anything.

SURMOUNT-5 (obesity, no diabetes) randomised 751 adults to maximum tolerated tirzepatide (10 or 15 mg) or maximum tolerated semaglutide (1.7 or 2.4 mg) for 72 weeks. Tirzepatide produced 20.2% weight loss against 13.7% - a 6.5-point absolute gap, or 47% more relative weight loss. Tirzepatide also won on waist circumference, systolic and diastolic blood pressure, HbA1c, fasting insulin, triglycerides, and HDL cholesterol.

SURPASS-2 (type 2 diabetes) randomised 1,879 adults on metformin to tirzepatide 5, 10, or 15 mg or semaglutide 1 mg for 40 weeks. HbA1c fell 2.01, 2.24, and 2.30 points against 1.86. Weight fell 7.6, 9.3, and 11.2 kg against 5.7 kg. On the composite target - HbA1c at or below 6.5%, at least 10% weight loss, no severe hypoglycaemia - 60% of the tirzepatide 15 mg group got there against 22% on semaglutide.

One caveat worth stating plainly, because Novo Nordisk is entitled to it: SURPASS-2 capped semaglutide at 1 mg, which was the highest approved dose at the time but is half of Wegovy’s ceiling. SURMOUNT-5 does not have that problem - it used semaglutide at its full obesity dose - which is why it is the more decisive of the two.

The number to plan around is not the number in the headline

Every trial reports its result two ways, and press releases quote the flattering one. The efficacy estimand asks what happens in people who took the drug as directed for the full duration. The treatment-regimen estimand asks what happened to everyone randomised, including those who stopped. The gap between them is the cost of real life.

It is not small. Orforglipron’s pivotal trial reads 12.4% by the first convention and 11.2% by the second. Oral semaglutide reads 16.6% and 13.6%. Anyone comparing a press-release number for one drug against a journal number for another is comparing nothing.

Then there is the clinic, which is harsher still. A propensity-matched study of more than 18,000 US patients published in JAMA Internal Medicine found 12-month weight change of -15.3% on tirzepatide and -8.3% on semaglutide. Those patients were mostly on the diabetes-labelled products at diabetes doses, so the absolute figures sit below the obesity trials by design. The instructive part is the other number in that paper: 55.9% of tirzepatide patients and 52.5% of semaglutide patients had stopped within twelve months.

That pattern repeats everywhere it has been measured. A Danish national cohort of 77,310 adults taking semaglutide for weight loss found 52% discontinued inside a year - 18% by three months, 31% by six, 42% by nine. Persistence is improving as prescribing matures, from 33.2% in 2021 to 60.9% in 2024, but the base rate is still that roughly half of people who start this class do not finish their first year.

The practical reading: the difference between these two molecules is smaller than the difference between taking one of them and stopping. Choose partly on which one you can actually stay on, at a price you can still pay in month fourteen.

What each drug costs per point of weight lost

Monthly price is the wrong unit. A cheaper drug that does less is not cheaper. So we priced each trial regimen - the actual titration schedule, at US manufacturer self-pay prices as of August 2026, one 28-day fill per month, over that trial’s own duration - and divided by the weight loss that trial reported on the treatment-regimen basis.

Tirzepatide is the most expensive course on the list and the second-best value on it, because the denominator moves faster than the numerator. Injectable semaglutide costs 20% less over the course and delivers 32% less weight loss, which is how a $459-per-point figure comes out of a cheaper drug.

Workings, so you can redo them: Zepbound $299 + $399 + (16 x $449) = $7,882. Wegovy injection 18 x $349 = $6,282. Wegovy pill $149 + $199 + $299 + (13 x $299) = $4,534. These are cash prices through LillyDirect and NovoCare and they exclude insurance entirely - with commercial coverage and a savings card, either drug can land near $25 a month, at which point the arithmetic above is irrelevant and the clinical questions are all that is left. Self-pay prices in this category have moved several times a year since 2024; check them before relying on them.

Where semaglutide is clearly ahead: proof of organ benefit

This is the part of the comparison that gets flattened into a footnote, and it is the part most likely to change a decision.

SELECT randomised 17,604 adults with cardiovascular disease and overweight or obesity, without diabetes, to semaglutide 2.4 mg or placebo. Semaglutide cut major adverse cardiovascular events - cardiovascular death, non-fatal heart attack, non-fatal stroke - by 20%. That is a superiority finding against nothing at all, which is the strongest form this evidence takes.

SURPASS-CVOT, published in December 2025, randomised more than 13,000 patients with type 2 diabetes and cardiovascular disease to tirzepatide or dulaglutide, an older GLP-1 drug, over 4.5 years. Tirzepatide was non-inferior, with an 8% lower rate of major events. That is a real, reassuring result and it is not the same claim. It establishes that tirzepatide is at least as good as a drug already known to help; it does not independently establish by how much it beats no treatment.

Semaglutide has also collected labels tirzepatide does not have: cardiovascular risk reduction in obesity (March 2024), MASH with moderate-to-advanced fibrosis (accelerated approval, August 2025), and diabetic kidney disease under the Ozempic brand. Tirzepatide holds the one label semaglutide does not: moderate-to-severe obstructive sleep apnea in adults with obesity, granted December 2024 and the first drug approval for that condition in any class.

So the honest framing is not "which drug is better" but "which endpoint is yours." Weight and blood sugar point to tirzepatide. Established cardiovascular disease, liver fibrosis, or kidney disease point to semaglutide, because that is where the evidence was actually generated. Sleep apnea points back to tirzepatide.

The tolerability result almost nobody predicts

The intuition is that a drug hitting two receptors should be harder to tolerate. SURMOUNT-5 found the opposite: gastrointestinal side effects drove 2.7% of tirzepatide patients off the drug and 5.6% of semaglutide patients, roughly a two-fold difference in the direction nobody bets on.

Both drugs produce nausea, constipation, diarrhoea, and vomiting, mostly mild to moderate and mostly clustered in the weeks after each dose increase rather than at steady state. The single largest lever on whether you tolerate either one is not which molecule you picked - it is how fast you climbed. Titration intervals exist because they work.

Your own data beats both trials

Every figure on this page is a group average. SURMOUNT-5 found a 6.5-point gap between the molecules, but the spread within each arm was far wider than the gap between them - some people on semaglutide outlost the tirzepatide median, and some people on tirzepatide barely moved. A trial tells you what happened to a population. It cannot tell you which end of that distribution you are on.

The only way to find out is to measure, and to measure the right things. Weight once a week at a fixed time, not daily. Dose and dose-change dates, because response is judged against the ladder rather than the calendar. Side effects with a severity and a date, because that is what tells you whether to hold or climb. Waist circumference monthly, since it moves when the scale stalls.

Zenday App is our top-ranked GLP-1 companion at 4.55/5, and it is the one built around exactly this: it holds the dose schedule, the food, and the side-effect days in one place and turns them into a week you can plan rather than absorb. Past a million users, designed with doctors, founded by a GLP-1 patient - and an independent 2026 study puts six-month weight loss at 2.4x medication alone. That multiplier is worth sitting with, because it is larger than the entire difference between the two drugs this article compares.

There is a second reason to log from day one, and it is the one the persistence data points at. Roughly half of people stop within a year, and side effects are the most common reason. Managing them is a tracking problem before it is a pharmacology problem - which is why we weight side-effect handling heavily in theGLP-1 app ranking, and why the apps that treat it as a core pillar rather than a symptom checklist score above the ones that do not.

What the evidence does not show

• No head-to-head against the newest doses.SURMOUNT-5 compared tirzepatide 15 mg against semaglutide 2.4 mg. Wegovy HD 7.2 mg, approved in March 2026, reached 20.7% in its own placebo-controlled trial - statistically indistinguishable from tirzepatide’s 20.2% in a different study. Nobody has run those two against each other, and until someone does, the efficacy gap between the molecules at their current ceilings is genuinely unsettled.
• No placebo-controlled cardiovascular outcome trial for tirzepatide. Absence of evidence, not evidence of absence - but you cannot cite what has not been run.
• No long-term comparative safety data. SURMOUNT-5 ran 72 weeks. These are drugs people are being told to take indefinitely.
• Nothing about compounded product. Every figure on this page comes from trials of pharmaceutical-grade drug at verified concentrations. None of it transfers to grey-market vials of unverified content.

How to choose

• Weight loss is the goal and cost is manageable: tirzepatide. It wins the only head-to-head that tested both at full obesity doses, and it costs less per point of result.
• You have established cardiovascular disease: semaglutide, on the strength of SELECT.
• You have MASH with fibrosis, or diabetic kidney disease: semaglutide. The labels exist for a reason.
• You have moderate-to-severe obstructive sleep apnea: tirzepatide, which is the only drug approved for it.
• You failed semaglutide on side effects: tirzepatide is a reasonable next move rather than a harder one, on the SURMOUNT-5 dropout data. See switching from semaglutide to tirzepatide.
• Needles are the barrier: neither, in injectable form. Compare Foundayo against tirzepatide, and note that the Wegovy pill reached 13.6% - close to the injection - at the price of an empty stomach and a 30-minute wait every morning.

Method and sources

Efficacy figures are the treatment-regimen or treatment-policy estimands from SURMOUNT-5, SURPASS-2, SURPASS-CVOT, SELECT, STEP UP, and OASIS-4 as published, not the press-release estimands, except where labelled otherwise. Real-world figures come from the JAMA Internal Medicine propensity-matched cohort and national persistence studies cited above. Prices are US manufacturer self-pay rates published by LillyDirect and NovoCare and read on 28 August 2026; they change frequently and are not the prices you will pay with insurance. We take no money from any manufacturer, run no affiliate links, and hold no position in either company. Corrections go to our contact page and we publish them.

If you are working with research-grade peptides rather than prescriptions, our dosing protocols and reconstitution calculator cover the arithmetic, and the GLP-1 app ranking covers the tracking tools.