Walk into any conversation about weight-loss medication and someone will tell you tirzepatide beats semaglutide, or that the difference is hype. Until recently both claims floated on indirect comparisons: one trial here, another there, different people, different years. In 2025 the guessing stopped, because someone finally ran the direct contest.
The trial that settled it
SURMOUNT-5 randomised 751 adults with obesity (but without diabetes) to the highest tolerated weekly amount of either tirzepatide or semaglutide, then followed them for 72 weeks. The headline number: average weight change of −20.2% on tirzepatide versus −13.7% on semaglutide. Waistlines told the same story, −18.4 cm against −13.0 cm. People on tirzepatide were also more likely to cross every meaningful threshold the trial set, from 10% all the way to 25% lost.
Context helps these numbers land. In semaglutide’s own landmark trial, STEP 1, the drug group lost 14.9% against 2.4% for placebo over 68 weeks. In tirzepatide’s SURMOUNT-1, the highest amount lost 20.9% against 3.1% for placebo. So the head-to-head did not reveal a different tirzepatide; it confirmed the earlier trials were roughly telling the truth, then put the gap between the two drugs at about six to seven percentage points.
Why one number is bigger
The short pharmacology: semaglutide mimics one gut hormone signal (GLP-1). Tirzepatide mimics two (GLP-1 plus GIP). The dual action suppresses appetite further in most people, which is the entire mechanism that matters here. Neither drug burns fat directly; both make eating less feel normal, and the weight follows the food.
Side effects rhymed. In both groups the common complaints were gut-related, mostly mild to moderate, mostly during the ramp-up phase while amounts were being raised. Nothing in the trial suggested the extra weight loss came with a different category of harm.
Three things people get wrong about this result
It was open-label. Everyone knew which drug they were getting. That is standard for trials where the pens look different, and randomisation still protects the comparison, but it is worth knowing before treating the numbers as physics.
Averages are not promises. −20.2% is a mean with a confidence interval, not a personal forecast. Some participants lost far more, some far less, and the trial says nothing about who you will be.
Weight is the measured thing, not health. The trial ran 72 weeks and measured kilos and centimetres. It did not measure heart attacks averted, years gained, or what happens in year five. Weight loss of this size almost certainly matters for health, but “almost certainly” is doing real work in that sentence.
The practical read
If the question is which drug moves the scale more on average, SURMOUNT-5 answers it: tirzepatide, by a margin that is hard to call small. If the question is which one a person should take, the trial is silent. Tolerability, cost, availability, medical history and a clinician’s judgment fill that gap, and none of them appear in a weight-change table.
As always on this site: the numbers came from the sources in the rail, not from us.
The trial answers ranking, not access. If you are in a country where neither drug is licensed and you are looking at research-use material sold by chemical name, the semaglutide listings are the same compound under a different supply chain, with none of the trial’s purity or cold-chain data attached to them. Which drug wins on the scale does not change. Whether what you hold is what was in the vial does.
