Orforglipron vs oral semaglutide in T2D

A 52-week randomised trial has answered a question the GLP-1 market has been arguing about for two years. If both drugs arrive as a daily tablet, does Eli Lilly's orforglipron lower blood sugar better than Novo Nordisk's oral semaglutide? In ACHIEVE-3, published in The Lancet in 2026, the answer on HbA1c was yes at every dose pairing tested^[1]. The answer on tolerability went the other way. Both halves of that result matter, and the headline only carries one of them.

What ACHIEVE-3 actually tested

ACHIEVE-3 was a phase 3, open-label, active-controlled trial run at 131 centres in Argentina, China, Japan, Mexico and the United States, registered as NCT06045221^[2]. It recruited 1,698 adults between September 2023 and August 2025. Everyone enrolled had type 2 diabetes that metformin at 1,500 mg per day or more had failed to bring under control, an HbA1c between 7.0% and 10.5%, and a BMI of at least 25^[1].

Participants were split four ways: orforglipron 12 mg (424 people), orforglipron 36 mg (423), oral semaglutide 7 mg (426) and oral semaglutide 14 mg (425). Each group had a lead-in period of up to four weeks, then 52 weeks of one tablet a day^[1].

The design point worth holding onto is what the trial set out to prove. Its primary objective was non-inferiority, with a margin of 0.3 percentage points of HbA1c. Superiority testing was prespecified but only unlocked once non-inferiority was met. Orforglipron cleared both bars^[1].

The HbA1c result

Baseline HbA1c averaged 8.3%. Under the treatment regimen estimand, which counts everyone randomised regardless of whether they stopped the drug or added another one, mean HbA1c change at week 52 was -1.71% on orforglipron 12 mg, -1.91% on orforglipron 36 mg, -1.23% on oral semaglutide 7 mg and -1.47% on oral semaglutide 14 mg^[1].

As head-to-head differences, orforglipron 12 mg beat semaglutide 7 mg by 0.48 percentage points (95% CI 0.31 to 0.65, p<0.0001), and orforglipron 36 mg beat semaglutide 14 mg by 0.44 points (95% CI 0.26 to 0.62, p<0.0001)^[1]. The comparison that drew most of the attention was the mismatched one. Orforglipron at its lower 12 mg dose still outperformed semaglutide at its higher 14 mg dose, by 0.24 points (95% CI 0.072 to 0.41, p=0.0050)^[1].

The ACP Journal Club at McMaster University selected the trial for its evidence digest in Annals of Internal Medicine in June 2026, and led with the same finding^[3].

The tolerability cost

Gastrointestinal side effects were the most frequent adverse events in every group, which is ordinary for this drug class. The rates were not even. Around 59% of the orforglipron 12 mg group and 58% of the 36 mg group reported a gastrointestinal event, against 37% on semaglutide 7 mg and 45% on semaglutide 14 mg^[1]. Most were mild or moderate in severity.

Discontinuation followed the same shape. Nine percent of the orforglipron 12 mg group and 10% of the 36 mg group stopped treatment because of an adverse event, roughly double the 4% and 5% recorded on the two semaglutide doses^[1]. Mean pulse rate also rose further on orforglipron, by 3.7 bpm at 12 mg and 4.7 bpm at 36 mg, against 1.0 and 1.5 bpm on semaglutide^[1].

Four deaths occurred during the trial, one in each orforglipron group and two in the semaglutide 7 mg group^[1]. Liver safety, a question that ended an earlier Lilly oral candidate, looks clean so far. A pooled analysis of 11,220 participants across seven phase 3 orforglipron trials found no case meeting drug-induced liver injury or Hy's Law criteria, with hepatic adverse events balanced against comparators^[4].

What the design cannot tell you

ACHIEVE-3 was open-label. Every participant knew which tablet they were swallowing, which is a real limitation once the outcomes include self-reported nausea and appetite. Eli Lilly funded the trial, and six of the eight named authors are Lilly employees and shareholders^[1].

The comparator doses also bound the conclusion. Oral semaglutide was tested at 7 mg and 14 mg, the strengths licensed for type 2 diabetes in Europe as Rybelsus^[5]. A 25 mg oral semaglutide dose exists, and it has been compared with orforglipron 36 mg in obesity through a Novo Nordisk funded indirect analysis that favoured semaglutide on both weight loss and discontinuation^[6]. No one has run orforglipron against high-dose oral semaglutide in diabetes, and an indirect comparison between two company-run trials is not the same evidence as a randomised head-to-head.

Fifty-two weeks is short for a diabetes medicine. There is no cardiovascular outcome data for orforglipron yet, and cardiovascular outcomes are the evidence semaglutide built its guideline position on.

What this means if you have type 2 diabetes

Orforglipron is a non-peptide small molecule rather than a modified hormone, which is why it was designed to be taken without the food and water restrictions oral semaglutide needs for absorption^[1]. For anyone whose HbA1c has not come down far enough on metformin, ACHIEVE-3 says a second oral option exists and that it lowered blood sugar more than the incumbent pill over a year. It also says you are more likely to feel sick on it and more likely to stop taking it. Which of those weighs heavier is an individual clinical judgement, not something a trial average decides for you.

Licensing, availability and brand names differ by country, and our semaglutide regulation page sets out where the oral form is authorised. This article is for educational and journalistic purposes and is not medical advice. Any decision about starting, switching or stopping a diabetes medicine belongs with your prescribing clinician.

What we don't yet know

Orforglipron reached patients first as a weight-loss medicine, where a 2026 pharmacotherapy review positioned it as probably better than liraglutide, broadly comparable to semaglutide, and behind tirzepatide^[7]. Its place in diabetes care is less settled. We do not know how the two drugs compare past 52 weeks, whether the HbA1c advantage survives real-world adherence given the higher dropout rate, or what a sustained pulse-rate increase means across years rather than months. Comparisons run by someone other than the two manufacturers would answer more than either company's own trial can.