ACHIEVE-3 topline: orforglipron beats oral semaglutide head to head

Eli Lilly announced topline results from ACHIEVE-3, the first head-to-head Phase 3 trial comparing two oral GLP-1 receptor agonists. In 1,698 adults with type 2 diabetes who were not well controlled on metformin, orforglipron lowered A1C and body weight more than oral semaglutide (sold as Rybelsus) over 52 weeks ^[2].

The trial was open-label, meaning participants and doctors knew which drug was being taken. People were randomly assigned in equal numbers to orforglipron 12 mg, orforglipron 36 mg, oral semaglutide 7 mg, or oral semaglutide 14 mg, at sites in the U.S., Argentina, China, Japan, Mexico and Puerto Rico. Semaglutide was given according to its approved label instructions, which include timing around food and water. Both drugs were started low and stepped up every four weeks ^[2].

What the numbers showed

Starting from an average A1C of 8.3%, A1C fell by 1.9 points with orforglipron 12 mg and 2.2 points with 36 mg, versus 1.1 points with oral semaglutide 7 mg and 1.4 points with 14 mg, using the "efficacy estimand" — an analysis that estimates results as if everyone had stayed on treatment without adding other diabetes drugs ^[2].

From an average starting weight of 97.0 kg (213.9 lbs), participants lost 6.7% (14.6 lbs) on orforglipron 12 mg and 9.2% (19.7 lbs) on 36 mg, compared with 3.7% (7.9 lbs) and 5.3% (11.0 lbs) on oral semaglutide 7 mg and 14 mg — a 73.6% greater relative weight loss at the top doses ^[2]. Near-normal blood sugar (A1C below 5.7%) was reached by 37.1% on orforglipron 36 mg versus 12.5% on oral semaglutide 14 mg ^[2].

A second analysis, the treatment-regimen estimand, which counts everyone regardless of whether they stayed on the drug, showed smaller but still significant gaps: A1C down 1.7% and 1.9% with orforglipron versus 1.2% and 1.5% with oral semaglutide, and weight down 6.1% and 8.2% versus 3.9% and 5.3% ^[2].

Side effects were mostly gastrointestinal and generally mild to moderate. The most common were nausea, diarrhea, vomiting, indigestion and decreased appetite ^[1]. More people stopped orforglipron because of side effects: 8.7% (12 mg) and 9.7% (36 mg), versus 4.5% and 4.9% for oral semaglutide ^[2]. Lilly noted the study was not powered to compare safety and tolerability between the drugs, and said no liver safety signal was seen with orforglipron ^[2].

Detailed results were later published in The Lancet. That release added that 85.4% of participants on orforglipron 36 mg reached A1C below 7%, versus 66.1% on oral semaglutide 14 mg ^[1]. Lead investigator Dr. Julio Rosenstock said differences appeared "as early as four weeks and sustained throughout the study" ^[3].

Why it matters for patients

Most GLP-1 comparisons rely on separate trials with different participants, which makes cross-study math unreliable. A head-to-head trial removes some of that guesswork for these two pills at these specific doses ^[2].

Several caveats matter. The semaglutide doses tested, 7 mg and 14 mg, are diabetes doses, not the higher doses studied for weight loss ^[1]. The trial enrolled people with type 2 diabetes on metformin, so results may not transfer to people without diabetes. The open-label design can influence reported side effects and behavior. Weight loss and some A1C thresholds were secondary endpoints, and a few comparisons were not statistically controlled for multiple testing ^[1][2]. And the higher discontinuation rate on orforglipron suggests the larger average benefit may come with tolerability trade-offs for some people ^[2].

Orforglipron is taken once daily with no restrictions on food or water timing, unlike oral semaglutide ^[1]. Cost, insurance coverage and pricing are not addressed in these sources.

What happens next

In the September 17, 2025 topline release, Lilly said full results would be presented at a medical meeting and published, and that it expected to file orforglipron for type 2 diabetes with global regulators in 2026 ^[2]. On February 26, 2026, Lilly said the ACHIEVE-3 results were published in The Lancet, that orforglipron had been submitted to regulators in more than 40 countries, that U.S. regulatory action for obesity was possible in the second quarter of 2026, and that the U.S. type 2 diabetes submission was planned later that year ^[1][3]. Detailed results from three remaining ACHIEVE registration trials are expected in 2026 ^[1].