# How Much Weight Will I Lose on a GLP-1? What Predicts It | Buy This GLP

Across 135,349 treated people, the share losing over 15% ran from 10% to 34% by brand | moderate response of 5% to 15% held steady at 40% to 42% across all four.

Publisher: Buy This GLP Editorial Team
Canonical: https://buythisglp.com/research/how-much-weight-will-i-lose-on-a-glp1
Updated: 2026-11-14
Published: 2026-09-17

[← Research](https://buythisglp.com/research)

Evidence

# How Much Weight Will I Lose on a GLP-1? What Predicts It

In a study of four GLP-1 brands, 10% to 34% of patients lost more than 15% of their starting weight during the first year. About four in ten lost 5% to 15%.

Editorial profile: [Carla Medina](https://buythisglp.com/authors/carla-medina)[Buy This GLP Editorial Team](https://buythisglp.com/about)September 20269 min readUpdated November 2026

Model discrimination (AUC), same data · predicting weight loss · 0.94 · predicting blood sugar control · 0.79 · Top predictors of weight loss: baseline BMI and body weight. · Bars start at 0.5, which is chance.

**Research, not medical advice.** Buy This GLP is independent and educational. Nothing here is a substitute for individualized care from a licensed clinician, and reading it creates no doctor–patient relationship. Talk to your own provider before starting, stopping, or changing any medication, peptide, or therapy. See our [methodology](https://buythisglp.com/methodology) and [terms](https://buythisglp.com/terms).

## Key takeaways

-   Researchers trained several machine learning models on real-world data to predict which GLP-1 users would achieve weight loss and which would achieve glycemic control.
-   Tree-based ensembles reached about 0.94 discrimination and 0.89-0.90 accuracy for weight loss.
-   For glycemic control the same methods managed about 0.79 discrimination and 0.73 accuracy.
-   The most influential predictors of weight response were baseline BMI and baseline body weight.

The evidence on [Zepbound weight loss at three months](https://buythisglp.com/research/how-much-weight-can-you-lose-on-zepbound-in-3-months) provides an early checkpoint. In a separate study of 135,349 people treated with semaglutide or tirzepatide, outcomes during the first year varied by brand: 10% to 34% lost more than 15% of their weight [\[2\]](https://buythisglp.com/research/how-much-weight-will-i-lose-on-a-glp1#ref-2). Across all four brands, 40% to 42% lost between 5% and 15%.

The [brand comparison](https://buythisglp.com/research/super-responders-and-the-brand-confound) explains why treatment groups can differ even when they use the same active ingredient. Ozempic and Wegovy contain semaglutide, and Mounjaro and Zepbound contain tirzepatide. The products have different approved uses and dose schedules.

The guide to [how long weight loss can take](https://buythisglp.com/research/how-long-does-it-take-for-a-glp1-to-work) puts early changes in context. In a post hoc analysis of two tirzepatide trials, participants who had lost at least 5% at week eight lost more weight by week 72. Those below 5% at week eight also achieved clinically meaningful weight loss as a group [\[3\]](https://buythisglp.com/research/how-much-weight-will-i-lose-on-a-glp1#ref-3).

A separate study tested seven machine learning models on existing patient records. For the strongest models, the area under the receiver operating characteristic curve was about 0.94 for weight loss and 0.79 for blood sugar control. Accuracy was about 89% to 90% for weight loss and 73% for blood sugar control [\[1\]](https://buythisglp.com/research/how-much-weight-will-i-lose-on-a-glp1#ref-1).

Look at what the weight model is using

Baseline BMI and body weight were the model’s strongest predictors of weight change. The researchers also classified outcomes using a threshold of about 5% weight loss from baseline. The reported score comes from internal testing. The study did not test whether the model predicts outcomes in a separate patient population [\[1\]](https://buythisglp.com/research/how-much-weight-will-i-lose-on-a-glp1#ref-1).

The [prediabetes screening analysis](https://buythisglp.com/research/should-you-take-a-glp1-for-prediabetes) explains why a performance measure needs context. In this study, diabetes duration, baseline HbA1c and use of sulfonylureas or insulin were among the strongest predictors of blood sugar outcomes. The researchers tested the models using ten-fold cross-validation within their dataset [\[1\]](https://buythisglp.com/research/how-much-weight-will-i-lose-on-a-glp1#ref-1).

There is also nothing here a buyer can act on. The predictors are baseline characteristics, so learning that your starting BMI predicts your response does not give you anything to change. The guide to [why some people lose less weight on a GLP-1](https://buythisglp.com/research/why-do-some-people-not-lose-weight-on-a-glp1) explains how factors during treatment can affect the result.

## Frequently asked

Can a model tell me whether the drug will work for me?

Not usefully yet. The weight-loss model scored highly but leaned most on baseline BMI and body weight, and no external validation on a separate population is described.

Why did the blood sugar model perform worse?

The models were less accurate for blood sugar control in this dataset. The study did not establish why the two outcomes differed or whether the gap would persist in another population.

Is there anything actionable in it?

Not for a patient. Every predictor is a baseline characteristic, so the model describes who is likely to respond rather than anything that could be changed.

## Sources

1.  \[1\] Abegaz TM, Frietze G (2026). Machine learning algorithms for predicting glycemic control and weight loss outcomes in GLP-1 receptor agonist users Frontiers in Artificial Intelligence. [PMID 42529241](https://pubmed.ncbi.nlm.nih.gov/42529241/)
2.  \[2\] Venkatakrishnan AJ, Murugadoss K, Soundararajan V (2026). Decoding the hallmarks of GLP-1RA weight-loss super-responders Biology Methods & Protocols. [PMID 42147968](https://pubmed.ncbi.nlm.nih.gov/42147968/)
3.  \[3\] Kokkinos A, et al. (2026). Tirzepatide Efficacy and Tolerability According to Early Weight Response: A Post Hoc Analysis of the SURMOUNT-1 and SURMOUNT-2 Trials Diabetes, Obesity and Metabolism. [PMID 42348366](https://pubmed.ncbi.nlm.nih.gov/42348366/)

## Where to get it

### The cheapest GLP-1

Every seller that publishes a price, ordered lowest first on the standing month-to-month rate, plus the question to ask before you pay any of them.

[See the board →](https://buythisglp.com/best/cheapest-glp1)

## More in Evidence

-   [
    
    50,678 adults with a BMI of 30 or more · Prescribed · 8.0% · Filled · 4.4% · Outlined bar runs to 10%. Medicaid had lower odds of both.
    
    ### Does Medicaid Cover Weight Loss Drugs? It Depends on Your State
    
    ](https://buythisglp.com/research/does-medicaid-cover-weight-loss-drugs)
-   [
    
    What the claim rests on · Sleep study · AHI 15 or more · Obesity · BMI 30 or more · Prescription · for sleep apnea · Part D claim · plan review · Trial thresholds shown. Your plan sets its approval requirements.
    
    ### Does Medicare Cover Zepbound for Sleep Apnea? The Diagnosis Is the Key
    
    ](https://buythisglp.com/research/does-medicare-cover-zepbound-for-sleep-apnea)
-   [
    
    Illustrative decline with a one-week half-life · Ozempic window: up to 5 days late · 100% · day 0 · 82% · day 2 · 67% · day 4 · 61% · day 5 · 50% · day 7 · A calculation from the label half-life, not a blood test.
    
    ### What Happens If You Miss a Dose of Ozempic? A Week-Long Half-Life Buys Time
    
    ](https://buythisglp.com/research/what-happens-if-you-miss-a-dose-of-ozempic)
