BGM0504 Lowered HbA1c by 2.48 Points. The Semaglutide Comparison Needs Context
A 64-person Phase 2 trial found a 2.48-point HbA1c drop at the highest BGM0504 dose. The signal is real, but claims that it beat semaglutide need context.
A small trial has put another dual-incretin peptide into the obesity and diabetes race. BGM0504, a once-weekly experimental drug from BrightGene, lowered HbA1c by 2.48 percentage points at its highest dose after 12 weeks at target dose. That is a striking result. It is also exactly the kind of result that becomes misleading when the headline outruns the study.
The paper appears in the August 2026 issue of Diabetes, Obesity and Metabolism. It gives BGM0504 something many early obesity candidates still lack: peer-reviewed, randomized human data in people with type 2 diabetes. The catch is scale. Only 64 participants were randomized across five groups, the semaglutide arm was open-label, and the study was not powered as a definitive head-to-head contest.
So the right question is not whether BGM0504 has already beaten semaglutide or tirzepatide. It has not established that. The useful question is whether its glucose, weight, pharmacokinetic, and safety signals are strong enough to justify larger trials. On that narrower test, the answer is yes.
What the trial actually tested
Investigators enrolled Chinese adults ages 18 to 65 with type 2 diabetes, HbA1c between 7.0% and 10.0%, and body mass index between 19.5 and 35.0 kg/m². Participants were either newly diagnosed or had stopped diabetes medication before screening. The average participant was about 50 years old, weighed 76.2 kg, and had a BMI of 27.8 kg/m².
The trial randomized participants to weekly BGM0504 at 5, 10, or 15 mg, placebo, or semaglutide 1 mg. Each BGM0504 group and the pooled placebo group were small, with roughly a dozen participants apiece; the semaglutide group had 16. Doses were gradually increased before participants spent 12 weeks at target dose. The primary endpoint was change in HbA1c. Fasting glucose, two-hour post-meal glucose, body weight, pharmacokinetics, adverse events, and antidrug antibodies were also tracked.
That design can detect a promising dose-response signal. It cannot reliably settle fine differences between two active drugs, especially when baseline characteristics vary across tiny groups. The 15 mg BGM0504 group began with a mean HbA1c of 8.3%, while the semaglutide group began at 7.8%. More room to improve can influence the size of an HbA1c drop.
The glucose result is the clearest signal
At week 12, mean HbA1c changed by minus 1.72 points with 5 mg BGM0504, minus 1.94 with 10 mg, and minus 2.48 with 15 mg. Semaglutide 1 mg produced a 1.43-point reduction. Placebo rose by 0.28 points. All BGM0504 doses were statistically different from placebo, and the 15 mg dose was statistically different from semaglutide in the authors’ analysis, with p=0.0327.
The supporting glucose measures moved in the same direction. Fasting plasma glucose fell by 2.05, 2.52, and 4.00 mmol/L across the ascending BGM0504 doses. Two-hour post-meal glucose fell by 7.18, 7.17, and 8.73 mmol/L. A consistent dose response across HbA1c, fasting glucose, and post-meal glucose makes the efficacy signal more credible than one favorable endpoint in isolation.
Still, “statistically different in this exploratory trial” is not the same as “clinically superior.” The semaglutide arm used 1 mg, a diabetes dose rather than the 2.4 mg obesity dose, and participants knew they were receiving semaglutide because that arm was open-label. The trial also used last observation carried forward for missing efficacy data, an older imputation method that can introduce bias. A properly powered, blinded active-comparator study would be needed before making a superiority claim.
The weight loss was notable, but the clock was short
The 5 mg group did not show a significant weight change. Mean weight fell by 4.83 kg with 10 mg and 8.25 kg with 15 mg. The placebo-adjusted percentage differences were minus 6.58% and minus 9.55%, respectively. The high-dose group also showed a greater percentage reduction than the semaglutide arm in the study analysis.
Losing roughly eight kilograms during a short diabetes trial deserves attention. But short studies can exaggerate the apparent pace of change, particularly while doses and appetite effects are still settling. They cannot tell us where weight loss plateaus, how much returns after discontinuation, or whether lean tissue is preserved. Nor can this study provide meaningful evidence about cardiovascular outcomes, kidney outcomes, pancreatitis, gallbladder disease, or rare adverse events.
BrightGene separately reported positive topline results from a 652-person Phase 3 obesity study in May 2026, including mean weight reductions of 14.3%, 17.3%, and 19.2% across the three doses versus 3.1% with placebo over 52 weeks. Those figures are encouraging, but they came from a company press release. Until the full dataset is peer-reviewed or posted with enough detail for independent scrutiny, it should be treated as sponsor-reported evidence, not as equivalent to the published Phase 2 paper.
Why BGM0504 is different from a simple GLP-1 drug
BGM0504 activates both the GLP-1 and GIP receptors, placing it in the same broad mechanistic class as tirzepatide. Both pathways can support glucose-dependent insulin secretion. GLP-1 signaling also reduces appetite and slows gastric emptying, while GIP co-activation may add metabolic effects that are not reproduced by pushing GLP-1 activity alone.
The design story is unusually specific. In a 2024 Scientific Reports paper, researchers used molecular-dynamics simulations to study how tirzepatide-like peptides interact with the two receptors. They moved the fatty-acid side-chain attachment away from a lysine at position 20, aiming to preserve a receptor interaction while retaining the long circulation time needed for weekly dosing. In laboratory and animal experiments, BGM0504 showed two- to threefold greater agonist activity than tirzepatide. That is a molecular and preclinical comparison, not proof that the drug works better in patients.
The human pharmacokinetics fit the weekly concept. In the Phase 2 diabetes study, peak concentrations appeared around 24 hours after dosing and the reported elimination half-life ranged from about 109 to 160 hours. Exposure generally rose with dose. A prior 40-person Phase 1 study in healthy volunteers also supported weekly administration and reported dose-related weight loss, but it was designed primarily to assess tolerability and drug handling rather than treatment efficacy.
The safety trade-off is already visible
Most adverse events were described as mild or moderate, but they were common. Across the whole trial, 56 of 64 participants experienced at least one treatment-emergent adverse event. Every participant in the 15 mg group reported one. Diarrhea was especially dose-sensitive: it affected 38.5% at 5 mg, 36.4% at 10 mg, and 75% at 15 mg, compared with 6.3% on semaglutide. Abdominal distension affected roughly one-quarter of each BGM0504 group.
No hypoglycemia was recorded. Two serious events occurred, a ruptured cerebral aneurysm in a participant with a known aneurysm and lower-limb pain in another participant. Investigators judged them unrelated or possibly unrelated to treatment. With groups this small, however, absence of a rare safety signal tells us very little.
Some participants developed antidrug antibodies after treatment, although no neutralizing antibodies were found and investigators did not identify an effect on efficacy or safety. That finding is worth monitoring because BGM0504 is a modified peptide intended for chronic use. Larger and longer trials are where uncommon immune reactions and durability questions become visible.
What comes next
BGM0504 now has a coherent development package: receptor-design work, Phase 1 pharmacokinetics, a randomized Phase 2 diabetes signal, sponsor-reported Phase 3 obesity results, and additional registered studies. ClinicalTrials.gov lists trials in obesity and type 2 diabetes, including a Phase 2 obesity study that uses tirzepatide as an active comparator. That direct comparison will be more informative than comparing percentages across unrelated trials.
Geography also matters. The published participants were all Chinese adults, and the Phase 3 topline report described a study conducted at 41 sites in China. That is not a flaw, but it limits how confidently the findings can be generalized to populations with different body composition, background therapies, diets, and patterns of diabetes care. Multi-regional data would strengthen any global development case.
There is no basis here for self-experimentation. BGM0504 remains investigational, dose titration was supervised, and high-dose gastrointestinal effects were frequent. Products sold online under an experimental compound name do not inherit the identity, purity, or safety controls of the trial drug.
The Oria take
BGM0504 is no longer just an interesting molecule. A 2.48-point HbA1c reduction, parallel improvement in fasting and post-meal glucose, and substantial high-dose weight loss are enough to take the program seriously. The result also reinforces a larger trend: the next incretin competition is not simply about making another GLP-1. It is about tuning multiple hormone signals in one long-acting peptide.
But the cleanest interpretation remains modest. This was a 64-person, sponsor-funded exploratory trial with an open-label semaglutide arm and only about a dozen people per BGM0504 dose. It provides a strong signal, not a verdict. The next decisive evidence will come from full Phase 3 publication, transparent safety tables, longer follow-up, and adequately powered active-comparator trials.
Sources
Evidence grade: B. Randomized, placebo-controlled, peer-reviewed human Phase 2 data show a consistent dose response, but the sample was small, treatment was short, the semaglutide arm was open-label, and confirmatory comparative evidence is still needed.
Medical disclaimer: This article is for educational purposes only and is not medical advice. BGM0504 is an investigational drug and is not established here as approved for personal use. Do not buy, use, or change any peptide or diabetes treatment based on this article. Discuss treatment decisions with a licensed clinician who knows your history. Oria exists to help readers separate emerging peptide research from marketing claims, not to replace medical care.
Related Compounds
Build Your Personalised Protocol
Turn your research into a personalised supplementation protocol with the BioStack Generator.