Metformin vs. Ozempic for Weight Loss (and PCOS): What Each Actually Does

The short answer

Metformin is a cheap, decades-old insulin sensitizer that moves the scale about 2% on average. Ozempic's molecule, semaglutide, is a GLP-1 drug that averaged about 15% weight loss at the higher Wegovy dose. For PCOS, metformin is the old standard, and the two are often prescribed together.

That is the whole comparison in one breath. The rest of this piece is about why the numbers are so far apart, what each one does for a cycle that has gone quiet, and why the answer changes if you are planning a baby.

Two drugs that are not the same kind of thing

Metformin is not a GLP-1, and it is not a weaker version of one. It is a biguanide: it mainly turns down how much glucose the liver releases and helps tissues answer insulin better. Most of the appetite change women notice on it is modest and comes partly from stomach upset in the early weeks.

Semaglutide copies a gut hormone. It slows how fast the stomach empties and acts on appetite centers in the brain, which is why the effect on hunger is so much louder. Ozempic is the brand approved for type 2 diabetes, up to 2 mg a week; Wegovy is the same molecule approved for weight, up to 2.4 mg a week (and now a 7.2 mg high dose and a daily tablet). When a friend says she is "on Ozempic for weight loss," the trial numbers that apply to her are usually the Wegovy ones.

Weight loss: about 2% vs about 15%

The best long-term metformin data come from the Diabetes Prevention Program, which randomized 3,234 adults with prediabetes (68% of them women) to metformin, an intensive lifestyle program or placebo. Metformin cut new diabetes by 31% compared with placebo; lifestyle change cut it by 58%¹. On the scale, people taking metformin lost an average of 2.06% of their weight during the blinded trial, against 0.02% on placebo, and the gap held at about 2% over roughly ten years for those who kept taking it².

There is one encouraging detail in the follow-up. Among people who lost at least 5% in the first year, those on metformin kept off an average of 6.2% between years 6 and 15, more than either the lifestyle or placebo groups³. So metformin's weight effect is small, but it lasts for the people it works for.

Semaglutide's number is a different order of magnitude. In STEP 1, adults on the 2.4 mg Wegovy dose lost an average of 14.9% over 68 weeks, compared with 2.4% on placebo⁴. Ozempic's diabetes doses are lower, and people with diabetes tend to lose less, so do not read 15% as a promise from a 1 mg pen.

PCOS: cycles, androgens and insulin

This is where metformin has history, and where GLP-1s have momentum.

A Monash meta-analysis pooled 12 randomized trials in 608 women with PCOS. Adding metformin to lifestyle change lowered BMI a little (by 0.73 kg/m²) and added about one more menstrual cycle over six months compared with lifestyle alone. Metformin on its own was no better than lifestyle for weight, though testosterone came out lower⁵.

Head-to-head, the newer drugs look stronger, on thin evidence. A 2026 meta-analysis of four randomized trials (218 women) found GLP-1 drugs lowered testosterone, DHEA-S and androstenedione more than metformin did, while flagging substantial heterogeneity and possible publication bias⁶. A 2025 randomized trial in 100 women with PCOS gave metformin alone or metformin plus 1 mg semaglutide for 16 weeks: the combination lost an average of 6.09 kg against 2.25 kg, more cycles came back, and natural pregnancy rates over the following months were 35% against 15%⁷.

Tirzepatide has almost no PCOS trial data yet. A 2026 UK cohort of more than 54,000 women prescribed it found that women with self-reported PCOS lost about as much as women without it, around 19% at ten months, though only 40 of them were weighed at that point⁸. We go deeper on the hormone side in GLP-1s and PCOS: the evidence and the hype.

What women want from all this is not only a number. In a 2026 qualitative study, 19 women with PCOS were interviewed about their diagnosis and treatment. Participants described both as emotionally demanding, shaped by too little information and too little joined-up support, and many reported managing PCOS mostly on their own, through medication, diet and informal information seeking⁹. That is the gap a metformin-or-Ozempic decision lands in, which is why it deserves a clinician who knows PCOS rather than a quick script. Programs such as Allara Health are built around the condition underneath the weight.

Taking them together

Plenty of women do. Metformin is a common first-line diabetes drug, and semaglutide was tested on top of it in its diabetes trials; the Ozempic label lists it as an add-on to metformin¹¹. Neither drug usually causes low blood sugar on its own, but stacking two stomach-active drugs can stack the stomach upset, so ask whether to hold metformin at a dose you already tolerate while the weekly shot is titrated up.

Two practical notes from the metformin label: it can lower vitamin B12 over time, so a periodic check is reasonable, and it tells prescribers to discuss unplanned pregnancy with premenopausal women, because metformin may restart ovulation in some women who were not ovulating¹⁰.

Pregnancy and trying to conceive: the key difference

This is the part most comparisons skip, and for many mothers it decides everything.

Metformin has a long record in pregnancy. Its label says published studies have not shown a clear link to major birth defects or miscarriage¹⁰. A 2026 analysis that pooled individual data from seven placebo-controlled trials (2,297 pregnancies) found it did not prevent gestational diabetes, but it was linked to slightly longer pregnancies and fewer preterm births, with more stomach side effects¹⁵. Whether to continue it once you are pregnant is a call to make with your OB.

Semaglutide is the opposite. Both the Ozempic and Wegovy labels tell women to stop at least two months before a planned pregnancy because the drug clears so slowly^11,12. If you are in a trying-to-conceive window, that two-month runway can matter more than the weight-loss gap. We cover the planning in trying to conceive on a GLP-1.

What it costs a family

Generic metformin is one of the cheapest prescriptions in the country. Walmart's pharmacy program, for one, advertises select generics from $4 for a 30-day supply and $10 for 90 days; ask the counter whether your strength qualifies¹⁴.

Ozempic is a different budget line. On NovoCare's own pharmacy page today, new self-pay patients pay $199 a month for the first two fills of 0.25 mg or 0.5 mg (an offer running through December 31, 2026), then $349 a month for 0.25 mg, 0.5 mg or 1 mg, and $499 for 2 mg. Wegovy pens follow the same $199-then-$349 pattern, and the Wegovy tablet starts at $149¹³. Over a first year at the lower doses, that is roughly $3,900 for semaglutide, against under $50 for metformin at a $4-a-month price. Our cheapest GLP-1 board and what a GLP-1 costs without insurance walk through the rest of the market.

Which one, for whom

If your goal is preventing diabetes or nudging a stalled cycle, and your budget is tight, metformin is a reasonable, well-studied place to start. If you need a 10% to 15% loss for your health, metformin alone is very unlikely to get you there, and semaglutide or tirzepatide is the realistic tool. If you are trying for a baby in the next few months, the drug that can stay on board is metformin. And if you are weighing a supplement instead, read berberine vs. Ozempic before you buy a bottle.