It depends entirely on what you are comparing it against, and the published answers split along that line rather than along any line about the drug itself. Against diet and exercise, and against liraglutide, semaglutide comes out cost-effective. Against bariatric surgery it does not. Every one of those verdicts is a modeled lifetime cost from a payer’s point of view, which is a different quantity from the monthly figure a seller quotes. It is also a different question from which seller is cheapest.
There is a question underneath every price comparison on this site, and it is not which seller is cheapest. It is whether the drug is worth its price at all. Health economists have a formal way of answering that, and a 2024 systematic review pooled seven studies that had done so [1].
What the review found
Semaglutide produced more quality-adjusted life years than the comparator anti-obesity drugs — phentermine-topiramate, phentermine, and naltrexone-bupropion — and it cost more, with the incremental cost-effectiveness ratio exceeding the willingness-to-pay threshold in some cases. Against diet and exercise, and against liraglutide, it came out cost-effective. Against sleeve gastrectomy, endoscopic sleeve gastroplasty and gastric bypass, it did not.
The modeled lifetime costs in class I obesity (BMI 33) give the shape of it: lifestyle intervention at $124,195, endoscopic sleeve gastroplasty at $126,732, sleeve gastrectomy at $139,971, and semaglutide at $370,776.
The number the review does not give
A systematic review pools conclusions rather than ratios, so it is worth reading one of the underlying models directly. A 30-year Markov model from a US third-party payer perspective estimated that semaglutide 2.4 mg improved QALYs by 0.138 to 0.925 against its comparators, at a cost of $3,254 to $25,086 more. That puts the incremental cost at $23,556 to $144,296 per QALY gained [2]. Against a willingness-to-pay threshold of $150,000 per QALY, that is cost-effective against every comparator it tested — and the top of the range clears the bar by under six thousand dollars.
Two caveats belong with it. The model’s own sensitivity analysis named treatment duration, what happens after discontinuation, and weight-rebound rates as the key drivers of the result. Which is to say the answer depends most on the things nobody can promise. The study was also funded by Novo Nordisk, and two of its authors were employees of the company. The sponsor was involved in the study design, the interpretation of the data, the writing and the decision to submit. That does not make the arithmetic wrong; it means the reader should know whose model it is.
Why surgery wins on this measure and not on every measure
A bariatric operation is a large cost once. A GLP-1 is a smaller cost indefinitely, and indefinitely is the word doing the work — a drug taken for life accumulates past a one-time procedure at some point, and these models find that point. That is an argument about money over decades, not about which is right for a particular person, and it says nothing about risk, recovery, reversibility or preference.
It does have one practical implication for anyone reading a seller’s page. The cost-effectiveness case for this drug class rests on continuing to take it, and roughly one in three people are still on therapy at a year. A course that stops early carries the cost without the modeled benefit.
Cost per person who actually responds
Cost per prescription is the wrong denominator, because a prescription that does not work still costs the full amount. A UK analysis divided annual pharmacy cost by the share of patients reaching a composite target, which produces a cost per patient who actually responds[3]. At the strictest target — HbA1c at or below 6.5%, weight loss of 15% or more, and no hypoglycemia — tirzepatide came in £5,650 to £9,462 lower per patient reaching it than semaglutide. The gap widened as the targets got stricter, and reversed at the loosest one, where semaglutide cost less per responder.
The comparator in that analysis is semaglutide 1 mg, the diabetes dose, not the 2.4 mg weight-management dose. A cost-per-responder ratio computed against a lower comparator dose flatters the winner, and the direction of the mismatch runs the same way as the result. The full reading of that analysis sets out what the composite endpoint does and does not contain.
What a prevented heart event costs
The cardiovascular case has a number attached to it. A meta-analysis pooled 21 randomized trials and 99,599 patients across eight GLP-1 drugs. Over a mean 2.4 years, major adverse cardiovascular events fell, with an incidence rate ratio of 0.87 and a number needed to treat of 66[4]. All-cause death gave an NNT of 121, cardiovascular death 170, and the evidence was graded high certainty. Harms ran the other way: gastrointestinal disorders rose 63% and gallbladder disorders 26%.
Does it pay for itself?
Not within the year it is bought. A national household survey covering 7,144 person-years, of which 275 were sustained GLP-1 use, found non-drug medical spending lower by $2,586 per person-year among users and acute-care spending lower by $2,019 [5]. Put the drug back into the total and spending was higher, with a coefficient of 0.3182; the marginal dollar figure is not in the published abstract and is not computed here. A same-year analysis also cannot see a saving that arrives later, which is the saving the lifetime models are built to find.
Where a model does find one, the verdict rests on a threshold somebody chose. A 30-year model of liraglutide against sitagliptin in China returned an incremental cost-effectiveness ratio of 267,985.96 RMB per QALY gained. The threshold it was measured against was three times that country’s 2025 GDP per capita, a convention rather than a measurement. Against it, the drug came out cost-effective in 63.6% of probabilistic iterations [6]. In the other 36.4% it did not. Move the threshold and the verdict moves with it, which is worth remembering every time the phrase appears without the words “against what” attached.
What this changes at the checkout
Two things. First, the spread between sellers is not noise at this scale. Published semaglutide figures on this market run from $76 to $449 a month across 407 sellers, and over a multi-year course that is a difference measured in tens of thousands of dollars. The full distribution shows where any quote sits in it. Second, a headline that is not the billed figure compounds over that same horizon, and a large share of this market publishes one.
If the decision is close, the comparison worth running is not one seller against another but the whole course against the alternatives — which is a conversation for a clinician rather than a price page. The cost calculator will at least put a real number on the drug half of it.
Price figures are computed from this site’s own records at build time, most recently read September 2026. The economic figures are cited and are modeled, not observed.