Innovent presented two Phase 3 readouts for mazdutide at the American Diabetes Association meeting on Sunday, and the one that will travel is the head-to-head. In a trial called DREAMS-3, 48.0 percent of patients taking mazdutide ↗ hit both of their treatment goals at once. On semaglutide ↗, the active ingredient in Ozempic and Wegovy, 21.0 percent did. More than twice as many people cleared the bar on the Chinese drug.
The combined goal was strict. A patient had to get blood sugar under control, meaning an HbA1c below 7 percent (the standard marker of well-managed diabetes), and lose at least 10 percent of body weight, both in the same person. DREAMS-3 randomized 349 Chinese adults with early type 2 diabetes and obesity to mazdutide 6 mg or semaglutide 1 mg for 32 weeks. The edge held on each piece separately too: weight loss of 10.29 percent versus 6.00 percent, and an HbA1c drop of 2.03 points versus 1.84. Innovent, presenting through Linong Ji of Peking University People's Hospital, called it the first multinational Phase 3 trial to put a glucagon and GLP-1 dual agonist directly against semaglutide in diabetes, and win. The company released the result in a head-to-head announcement ↗ timed to the meeting.
The obesity number
The same weekend brought GLORY-2, mazdutide's obesity trial, published in JAMA ↗. It enrolled 461 Chinese adults with obesity, gave them either a 9 mg weekly dose or a placebo for 60 weeks, and measured weight. The drug group lost 16.65 percent of body weight on average. The placebo group lost 1.50 percent. The gap between the two, 15.15 percentage points, is the number that matters, and it is large for a once-weekly injection. Innovent's own materials cite up to 20.1 percent for patients who stayed on treatment the whole way, an on-treatment figure that runs ahead of the trial-wide average. Either way, 84.3 percent of the drug group lost at least 5 percent of their weight, against 33.1 percent on placebo, and the weight curve had not flattened by week 60.
Why the second receptor matters
What separates mazdutide from semaglutide is what it binds. Semaglutide hits one target, the GLP-1 receptor ↗. Mazdutide hits that one and the glucagon receptor ↗ as well. Glucagon is the hormone most people know as the one that raises blood sugar, which sounds like the wrong thing to switch on in a diabetes drug. In this pairing it appears to lift energy expenditure and pull on liver fat while the GLP-1 arm handles appetite and glucose. peptidemodel hosts mazdutide as a computational card ↗ against the glucagon receptor. The platform covered a different glucagon and GLP-1 dual, Boehringer's survodutide ↗, on Sunday for the same reason: the glucagon arm is the bet these molecules are making.
The catch is the one that follows the whole class. More than half the mazdutide group in GLORY-2 vomited (53.1 percent, against 1.3 percent on placebo), nearly half had nausea, and 39.4 percent had diarrhea. Most of it was mild to moderate, and only 2.9 percent of patients quit the trial over side effects, but the gastrointestinal load is real and it sits higher than some Western trials of this class report.
Mazdutide is already approved in China, where the regulator cleared it in 2025, and Innovent is developing it with Eli Lilly. The DREAMS-3 result is the part to watch. Semaglutide has lost a head-to-head before, but to tirzepatide, Lilly's own dual agonist. A second molecule beating it, from a different company on a different mechanism, says the single-receptor lead that defined the first wave of these drugs is closing.