The Pill Duel: Orforglipron Beats Oral Semaglutide Head to Head

Most drug trials are fought against a sugar pill, which tells you a drug works and almost nothing about whether it is the one to pick. ACHIEVE-3 is the rarer and more useful animal: a head to head. Two oral GLP-1 drugs, one clear winner on the numbers, a real price paid in side effects, and one important asterisk on who ran the race.

How the race was set up

The trial enrolled 1,698 adults with type 2 diabetes not controlled on metformin, at 131 sites in Argentina, China, Japan, Mexico and the United States. Entry required an HbA1c between 7.0 and 10.5 percent and a body mass index of at least 25. On average, people started at an HbA1c of 8.3 percent and a weight of 97 kilograms, about 214 pounds.

Participants were randomized in four equal groups: a lower or a higher dose of orforglipron, Lilly's small molecule pill, or a lower or a higher dose of oral semaglutide, the Novo Nordisk pill sold as Rybelsus. Doses were stepped up every four weeks, and treatment lasted 52 weeks. Oral semaglutide was taken according to its label, which comes with rules about food and water; orforglipron has none. The trial was open label, so everyone knew which pill they were on.

The formal goal was modest: show that orforglipron was no worse than oral semaglutide by more than 0.3 HbA1c points, then test for superiority. The primary analysis counted every randomized person, including those who stopped the pill or added another drug.

The result, in absolute terms

Orforglipron cleared both hurdles at both doses. Top dose against top dose, in the primary analysis, HbA1c fell 1.91 points on orforglipron and 1.47 on oral semaglutide: a gap of 0.44 points, with a 95 percent confidence interval of 0.26 to 0.62. Even the lower orforglipron dose beat the higher semaglutide dose, by 0.24 points.

You may have seen bigger numbers, 2.2 versus 1.4. Those come from the "efficacy estimand," a supportive analysis that asks what would have happened had everyone stayed on their assigned pill without rescue drugs. Lilly's first announcement led with it. Both are legitimate, but the one closer to real clinics, where people quit pills, shrinks the gap from about 0.8 points to under 0.5.

Weight moved the same way. In the same all comers analysis, top dose orforglipron cut body weight by 8.2 percent against 5.3 percent, roughly 8.1 versus 5.2 kilograms. On the threshold clinicians use most, an HbA1c under 7 percent, 75.8 percent of the orforglipron group got there against 63.8 percent on oral semaglutide. By my arithmetic, that 12 point gap means about one extra person reaching target for every eight treated with the winner instead of the runner up.

So the win is real, and smaller than the headline version. Half a point of HbA1c between two active drugs moves formularies at the margin; it does not end the argument.

The price: stomach and pulse

The Lancet paper is plainer about tolerability than the press releases. Gastrointestinal adverse events, mostly mild to moderate, hit 58 percent of people on top dose orforglipron against 45 percent on top dose semaglutide. About 10 percent of the top dose orforglipron group stopped treatment because of side effects, double the 5 percent on semaglutide. Average pulse rose 4.7 beats per minute on top dose orforglipron versus 1.5 on semaglutide. Four people died, one in each orforglipron group and two in the lower dose semaglutide group, too few to read anything into.

Lilly itself noted that the study was not powered to compare tolerability between the drugs, and that no liver safety signal appeared for orforglipron. So the efficacy comparison is solid; the side effect comparison is suggestive.

The asterisk

ACHIEVE-3 is Eli Lilly's trial, and orforglipron is Eli Lilly's drug. Lilly funded it, several Lancet authors are Lilly employees, and the sponsor chose the doses, the comparator, the population and the endpoints. That doesn't contaminate the data, and peer review adds a check on the methods. It does mean the correct reading is "orforglipron won the race its maker designed," which is still real information, and less than "orforglipron is simply better," which would need an independent rematch.

The comparator matters most. Oral semaglutide ran at the doses approved for diabetes in the United States; a higher pill dose exists but is not approved there for diabetes. In September 2026, Lilly presented an indirect comparison against that higher dose that still favored orforglipron, by smaller margins: 1.5 percent more weight loss and 0.3 points more HbA1c reduction. Indirect comparisons stitch separate trials together, so treat that as a hint, not a verdict.

A head to head against a real drug beats ten victories against placebo. A head to head run by one of the fighters beats nine of them.

Where orforglipron stands in October 2026

On April 1, 2026, the FDA approved orforglipron as Foundayo, 50 days after filing under its national priority voucher pilot, but for weight management in adults with obesity, or overweight with at least one weight related condition. It is not approved in the US for type 2 diabetes, the disease ACHIEVE-3 studied. Lilly says it submitted that application in the second quarter of 2026; I found no FDA decision on it as of this writing. In the United Kingdom, the MHRA authorized Foundayo on August 10, 2026, for both weight management and type 2 diabetes.

If you compare documents, note that Lilly's newer releases label the same ACHIEVE-3 groups with the strengths of the marketed tablets. Same trial, different labels. This article says "lower" and "higher" on purpose: dosing belongs with a prescriber.

What comes next

A year is short for a lifelong disease, and ACHIEVE-3 measured blood sugar and weight, not heart attacks, strokes or kidney failure. Lilly's ACHIEVE-4 trial, against insulin rather than semaglutide, reported cardiovascular safety results in September 2026, and a dedicated outcomes trial, ATTAIN-OUTCOMES, is still running. Nor does this trial say what happens to muscle as the weight comes off, a question we followed in the race to keep muscle on GLP-1s.

What to watch: the FDA's call on the diabetes indication, whether independent guidelines and payers treat a gap under half a point as decisive, and whether real world dropout tracks the trial's doubled discontinuation rate. Then the checklist in how to read a health study in 5 minutes applies directly. Efficacy wins trials; tolerability over years wins patients.