Most coverage of this topic treats GLP-1 agonists as one thing. They are not. The drugs differ in molecule, in receptor targets, in dose ceilings, and above all in how much anyone has actually studied them against alcohol. This article is part of our GLP-1 and Alcohol hub, and it goes drug by drug so you can see where the evidence stops and the extrapolation starts.
The short version: nearly all the human alcohol research is on semaglutide and exenatide. Everything else in the class is inference from the same mechanism, which is reasonable but is not the same as data.
# Semaglutide (Ozempic, Wegovy, Rybelsus)
The best-studied by a wide margin.
Semaglutide has the large electronic-health-record analyses showing lower rates of alcohol use disorder diagnosis and recurrence, and it has the small randomized trial published in JAMA Psychiatry in 2025 reporting reduced craving and fewer drinks per drinking day. It also carries the overwhelming majority of the patient reports, partly because it is the most widely prescribed.
If someone says “the GLP-1 alcohol research”, they almost always mean semaglutide. It has two articles of its own: does Ozempic reduce alcohol cravings for the diabetes dose, and Wegovy and alcohol cravings for the higher weight-management dose.
# Exenatide (Byetta, Bydureon)
The one with a proper trial that mostly did not work.
An older drug, and the subject of a randomized controlled trial published in JCI Insight in 2022 in patients with alcohol use disorder. Across the whole sample it did not reduce heavy drinking days. In a subgroup of participants with obesity, it did.
This result gets skipped in enthusiastic coverage, and it should not be. It is the most rigorous single trial in the field and its headline finding was negative. The honest reading is that the effect may depend on who is taking it, which fits the pattern that most positive findings come from populations with obesity or diabetes.
# Tirzepatide (Mounjaro, Zepbound)
Strong patient reports, thin formal evidence.
Tirzepatide is a dual agonist: it acts on the GIP receptor as well as GLP-1. That makes it pharmacologically distinct from everything else here, and it is more potent for weight loss.
Alcohol-specific research is limited. Some survey and self-report work, including a study published in Scientific Reports in 2023, found reduced drinking among people taking semaglutide or tirzepatide. Self-reported survey data collected from people who know what drug they are on is a weak design, and it should be read as a reason to run a trial rather than as a result.
Whether the GIP component changes the alcohol picture is unknown. Assuming tirzepatide behaves like semaglutide because both are “GLP-1 drugs” is an assumption, not a finding. The full picture, including the widespread belief that it is better tolerated than semaglutide, is in Mounjaro and alcohol cravings.
# Liraglutide (Victoza, Saxenda)
Some human work, mostly not about drinking.
Liraglutide is an older daily injection with a substantial research history, including brain-imaging work on food reward that supports the central-action part of the mechanism argument. Its alcohol-specific human evidence is thin. See Saxenda, Victoza and alcohol cravings.
# Dulaglutide (Trulicity)
Almost nothing specific to alcohol.
Dulaglutide is widely prescribed for diabetes and has little alcohol-focused research attached to it. A secondary analysis of a cardiovascular outcomes trial has been cited as hinting at reduced alcohol use, but a secondary analysis of a trial designed for a different question is a hypothesis, not a conclusion. See Trulicity and alcohol cravings.
# What this comparison actually tells you
Laid out side by side, three things stand out.
The evidence is concentrated, not distributed. Two molecules carry nearly all of it. The rest of the class is riding on a shared mechanism.
The most rigorous trial was largely negative. The exenatide result is the strongest reason for caution in the whole field, and it points at the same caveat as everything else: the positive signals cluster in people with obesity or diabetes.
Nobody has compared them against each other. There is no head-to-head trial of any two GLP-1 drugs for alcohol. Anyone ranking them for this purpose is ranking by prescribing popularity or by potency for weight loss, neither of which is the same question.
# Why the class-effect assumption might be wrong
Reading “GLP-1 agonists reduce alcohol cravings” invites the conclusion that any drug in the family will do it. That may turn out to be true. It is currently an assumption, and there are concrete reasons to hold it loosely.
They differ in how well they reach the brain. The alcohol effect is thought to involve receptors in the reward circuitry, which means a molecule has to act centrally rather than only on the gut and pancreas. These drugs are not identical in that respect, and a drug that works beautifully for blood sugar without much central action would not necessarily do anything for drinking.
They differ in half-life and dosing rhythm. A daily injection, a weekly injection and a daily tablet produce quite different concentration curves. If the effect depends on sustained receptor occupancy, that shape could matter.
Tirzepatide is not only a GLP-1 drug. Its GIP activity is a genuine pharmacological difference, and nobody knows whether that adds to the alcohol effect, subtracts from it, or does nothing.
The one trial designed to test the class hypothesis was mostly negative. Exenatide is a GLP-1 agonist, and in the population studied it did not reduce heavy drinking days. If this were a clean class effect, that result is harder to explain.
The safest reading is that semaglutide has evidence, the class has a hypothesis, and those are not the same claim.
# A note on compounded semaglutide
A large amount of semaglutide in circulation did not come from the licensed manufacturers. Shortages created a market in compounded versions, sold through telehealth services and less reputable channels, and that market has not gone away.
This matters for the alcohol question specifically. If the effect is dose-related, as covered in GLP-1 doses and alcohol cravings, then knowing your actual dose is the whole game. With a compounded product of uncertain concentration, you do not know it. Some are sold in different salt forms whose equivalence to the licensed product is not established.
Anyone trying to reason about their own response, and certainly anyone reporting it to a clinician, needs to know what they are actually taking. A product of uncertain provenance makes both the safety picture and the self-observation unreliable.
# What none of these drugs are
Worth stating clearly, because the framing in popular coverage often blurs it.
None of these drugs treat alcohol withdrawal. Someone physically dependent on alcohol who stops suddenly faces a genuine medical risk, and no GLP-1 agonist addresses that. Withdrawal management is a clinical matter, sometimes an urgent one.
None of them are a substitute for the rest of treatment either. The best evidence in alcohol care supports medication combined with some form of psychological or peer support, not medication alone. Whatever these drugs turn out to do, they will do it as one component.
And none of them are licensed for this. Every use described in this article is off-label, which means a prescriber can legally do it but is going beyond what the regulator has assessed.
# Choosing on this basis is the wrong way round
It is worth saying plainly: switching GLP-1 drugs to chase an alcohol effect is not something the current evidence supports, and it is not a decision to make from a comparison table.
The drug someone is on is chosen for their diabetes or their weight, by a clinician weighing their history, tolerance, and what is funded where they live. The alcohol question sits on top of that, and for now it sits there as an observation to monitor rather than a reason to change. Dose is more likely to be the relevant variable than brand, and that is covered in GLP-1 doses and alcohol cravings.
If reducing drinking is the actual goal, the licensed options exist and have the trial history this class does not: see GLP-1 vs naltrexone for alcohol cravings.
# How AlcoLog helps
Whichever drug you are on, the useful contribution you can make to your own answer is a record.
AlcoLog tracks medication doses alongside every logged drink, so you can see your own drinking against your own dose history rather than against a study population that may not resemble you. That matters here more than usual, because the research skews heavily toward people with obesity or diabetes, and if you do not fit that description the published findings may simply not apply to you.
The app does not fold medication into your AlcoScore. It records what you took and when, keeps it on your device with no account required, and exports it when you want to show a clinician.
