A monthly price tells a buyer which seller is cheaper. It cannot tell them which drug is cheaper, because the two molecules do not remove the same amount of weight for the money. The unit that would settle it is dollars per pound, and almost nobody publishes it — building it means holding a trial endpoint and a subscription invoice in the same hand, and the answer flatters neither number. The spread of outcomes underneath the averages used here is in the weight-loss tool.
Three inputs, each of them published
The first is a starting weight. Every percentage below is applied to 231 pounds, which is SURMOUNT-1’s own mean baseline body weight of 104.8 kg.[2] One weight for both molecules is what keeps the comparison honest; using each trial’s separate baseline would quietly change the denominator between rows.
The second is a loss. STEP 1 randomized 1,961 adults without diabetes to 68 weeks of semaglutide 2.4 mg or placebo, and reported a mean weight change of −14.9% against −2.4% on placebo.[1] SURMOUNT-1 randomized 2,539 adults to 72 weeks of tirzepatide at 5, 10 or 15 mg, or placebo, and reported −15.0%, −19.5% and −20.9% against −3.1%.[2]
The third is a price. The median advertised monthly figure across the 459 compounded semaglutide injection prices recorded on this site is $175; across the 415 compounded tirzepatide injection prices it is $240. Both are what sellers publish rather than what a walked checkout produced, and the structures that move them are in the annual-cost article. The trials, it should be said plainly, studied the approved products. Compounded drugs are not FDA-approved, and the agency does not verify their safety, effectiveness or quality before they are dispensed.
Semaglutide: about $80 a pound
STEP 1 ran 68 weeks, which is 15.64 calendar months rather than a year. At $175 a month that is $2,737 of medication to reach the published endpoint. A loss of 14.9% on 231 pounds is 34.4 pounds. Dividing gives $80 a pound.
The trial offers its own check on that conversion. STEP 1 also reported the change in absolute terms, at −15.3 kg against −2.6 kg on placebo.[1] That is 33.7 pounds, seven-tenths of a pound below the percentage route, and it moves the answer from $80 to $81. The arithmetic is not resting on the conversion.
Tirzepatide: about $82 a pound, at 37% more a month
SURMOUNT-1 ran 72 weeks, or 16.56 months, so the top dose costs $3,973 at the median $240. A loss of 20.9% on 231 pounds is 48.3 pounds, which is $82 a pound.
The monthly figures differ by 37%. The per-pound figures differ by about 2%, and in the direction that makes the more expensive drug marginally worse value rather than better. Almost the entire monthly premium is bought back by the extra 14 pounds, and nothing is left over. That is the whole finding, and it is invisible from a board sorted on monthly price.
The placebo arms were losing weight too
Both trials gave every participant lifestyle intervention, and both placebo groups lost weight on it — 2.4% in STEP 1, 3.1% in SURMOUNT-1.[1][2] A buyer paying for a drug is paying for the difference the drug made, not for the total on the scale.
STEP 1’s estimated treatment difference was −12.4 percentage points (95% CI, −13.4 to −11.5), which is 28.6 pounds at this weight and $96 a pound.[1] SURMOUNT-1’s top-dose gap of 17.8 points is 41.1 pounds and $97 a pound. The two molecules are, on this arithmetic, indistinguishable, and both are about 20% dearer than the gross figure suggests.
Inside one molecule, the dose decides the value
Where a seller holds one price across every rung — and a sizeable part of this roster does — the cost per pound moves only with the loss. At SURMOUNT-1’s figures and $240 a month, 5 mg costs $115 a pound, 10 mg costs $88, and 15 mg costs $82.
So the cheapest rung on a flat plan is the worst value on it, by 40% against the top. That inverts the usual reading of a dose ladder, where the low rungs look like the affordable part. They are only affordable per month; per pound they are the expensive part, which is a different argument for finishing a titration than the one the titration article makes on tolerability grounds.
Which seller you pick moves it far more than which drug
The medians above hide a spread that dwarfs the molecule comparison. The 10th percentile of the compounded semaglutide injection prices here is $99 a month and the 90th is $280. Run the same 68 weeks at each. At $99 the course costs $1,548, which is $45 a pound. At $280 it costs $4,379, which is $127 a pound.
That is a 2.8-fold range inside one molecule, against a 2% difference between the two molecules at their medians. Anyone treating the choice of drug as the cost decision has picked the smaller of the two variables by a factor of more than a hundred. Why one molecule carries that spread at all is worked through in the price-variation article.
A year of money does not buy a year of trial
Twelve months at the median prices is $2,100 for semaglutide and $2,880 for tirzepatide. Neither figure reaches a published endpoint. STEP 1’s result belongs to week 68 and SURMOUNT-1’s to week 72, so a calendar year of payments covers roughly three-quarters of the treatment that produced the numbers being quoted at it.
Anyone budgeting a year against a trial percentage is therefore short by about four months of medication. That is not a hidden fee; it is the difference between a marketing horizon and a protocol, and it is the reason every figure above is computed over the trial’s own length.
Stopping resets the denominator
STEP 1 carried an off-treatment extension. A subset of 327 participants who completed 68 weeks was followed for a further year after both semaglutide and the lifestyle intervention were withdrawn. On treatment, that subset had lost a mean of 17.3%. One year later they had regained 11.6 percentage points, leaving a net loss from baseline of 5.6%.[3]
Run the same division on what remains. A 5.6% net loss on 231 pounds is 12.9 pounds, against $2,737 already spent, which is $212 a pound — 2.6 times the figure at the endpoint. Nothing about the drug changed; the numerator simply stopped being spread across pounds that stayed off. What withdrawal looks like in the randomized trials designed to measure it is in the article on stopping.
That matters because stopping is common. A 2026 national survey of 440 United States adults who had started and then discontinued a GLP-1 found 54.6% stopped within six months and 79.7% within twelve. The reason cited most often was cost, at 36.1%, ahead of side effects at 33.3%.[4] The arithmetic in this article assumes a course run to the trial’s end, and four out of five people in that survey did not reach it.
What the formal literature does with the same question
Health economics asks this in cost per quality-adjusted life-year rather than cost per pound, and the results do not line up behind either molecule. A 2026 systematic review for the American College of Physicians, restricted to non-industry-sponsored United States evaluations, found none of 42 pairwise comparisons supported by high-certainty evidence. Among the moderate-certainty results, tirzepatide was high value against lifestyle modification, while semaglutide was low value against two much older drugs, naltrexone-bupropion and phentermine-topiramate.[5] Its stated conclusion is that study quality limits what can be concluded at all.
The comparison that does favor tirzepatide outright — a simulation on SURMOUNT-5 data estimating lifetime savings of $41,688 per patient and 0.506 additional quality-adjusted life-years against semaglutide — was authored by employees of the company that sells tirzepatide.[6] That does not make it wrong, and it is the reason the ACP review excluded industry-sponsored work by design.
What this arithmetic is not
It is not a projection for any individual. Trial means sit on top of wide distributions, and a person at the median of a price table is not thereby at the median of an outcome table. It also prices only medication: no labs, no shipping, no membership charged on its own clock, and no consequence of a plan that re-prices as the dose climbs.
And it assigns no value to the outcomes that are not weight. Both programs measured cardiometabolic endpoints alongside the scale, and a pound is a poor unit for those. What this division does is remove one specific illusion — that the cheaper monthly figure is the cheaper drug — using published numbers on both sides. The prices behind it are recorded in the seller write-ups.