Fake peptides: the five failure types seen in published lab tests, and the test that catches each

Fake peptides: five ways a vial fails a lab test

Published lab tests show five ways a peptide vial fails. The vial holds a different compound. It holds no detectable peptide. It holds too little or too much peptide for its label. The peptide is there but impure. Or the vial carries contamination, such as endotoxin. Each failure needs a different test, and only the first two are fakes in the strict sense.

Failure type	What the lab finds	Test that catches it	Published example
Wrong compound	The mass of a different molecule	Identity by LC-MS; blind GLP-1 identification	A vial sold as one GLP-1 peptide that holds another, such as tirzepatide for retatrutide
No peptide	No signal for the labelled peptide	Identity; content per vial	92 “non-detect” entries (no peptide, wrong peptide or a missing blend component) in one lab’s log, May 2026
Low or high content	mg per vial far from the label	Content per vial against a reference standard	Semaglutide vials 29% to 39% above the label (2024 study)
Low purity	The main peak is a smaller share of UV peak area	Purity by HPLC	142 MOTS-c entries below one lab’s 98.9% line (2026)
Contamination	Endotoxin in EU per mg; microbial growth	Endotoxin (LAL); Sterility (USP <71>, 14 days)	One of three test-bought samples at 8.95 EU/mg (2024 study)

Where the published failure data come from

No public data set breaks down every failure in the market by type. Four published sources each show part of the picture.

Source	What it covers	Size and dates	What it cannot tell you
Finnrick public data	Pass or fail for submitted and self-bought samples	11,999 published tests, latest dated 25 September 2026	Why each sample failed
Vanguard Laboratory failed-samples log, analysed by Krysia’s Substack	Failed samples only, with notes on the failure	2,106 entries, 17 February to 27 July 2026	How many samples passed
Ashraf and colleagues, JAMA Network Open	Test purchases of semaglutide from online sellers without a prescription	6 ordered, 3 delivered, August 2023 to March 2024	Anything about other products

Finnrick wrote in February 2026, after more than 7,000 samples, that “roughly one in three fail basic expectations”. Its default standard fails a sample for a wrong identity, purity under 99.5%, or a fill more than 35% above or below the label. So its fail count mixes fakes with vials of the right peptide that missed a quality line. The guide to peptide fail rates ranks products by fail share.

Krysia’s analysis of the Vanguard log breaks failures down by type. Its author warns that the log is “not its complete testing record”. It shows that recorded failures rose, from 337 in March to 594 in June. It cannot show whether the share of samples failing rose.

Wrong compound: the right label, a different peptide

A wrong-compound vial holds a real peptide, but not the one on the label. GLP-1 peptides show the risk: retatrutide, tirzepatide, semaglutide and cagrilintide are sold side by side and are all long, similar peptides. The guide to blind GLP-1 identification walks through a worked example.

A swap is easiest to catch by mass. An identity test by retention time compares when the sample leaves the HPLC column with a reference standard. ICH Q6A, the international guideline on specifications, says retention time alone “is not regarded as being specific”. LC-MS measures the mass of the molecule, and retatrutide and tirzepatide differ by 82 Da on molecules of about 4,700 to 4,800 Da. Our guide to LC-MS identity explains the difference.

A failed identity test does not always name the substitute. Finnrick’s help page says an identity fail means the chromatographic profile did not match the reference standard; it “does not tell you what the sample contains”. If you need the actual compound named, a blind GLP-1 identification asks the lab to name which GLP-1 family peptide is in the vial, from a stated list of candidates, without being told the label.

Names can also cause a mismatch. “CJC-1295” covers two molecules, with and without DAC, and labs list them as separate tests. A report is only useful when it states which molecule it tested for.

No peptide: empty vials and “faketide”

The most direct fake is a vial with no detectable peptide. In the Vanguard log, “non-detect” entries numbered 92 in May 2026, 67 in June and 77 by 27 July. Thirty non-detect entries were labelled retatrutide.

Non-detect is broader than empty. The article defines it as no peptide, the wrong peptide, or, in a blend, one expected compound not found. So these counts are an upper bound on empty vials, not a count of them.

“Faketide” is the lab’s internal name, reported in the same article, for a recurring unknown peak seen in several zero-peptide samples. The peak was not mannitol, a bulking agent, and matched no peptide in the lab’s library. On Reddit and peptide forums, buyers use the word more loosely, for fakes that could pass a lab that checks identity by retention time alone. A mass result tells the two apart: a compound that is not the labelled peptide does not give that peptide’s mass.

Heat is sometimes offered as an explanation for zero-peptide results. In the same article, the lab’s view was that peptides stay detectable after controlled heat exposure, and that ordinary transport temperatures would not make the peptide disappear completely. The guide on peptide degradation covers what storage does to results.

The test that catches an empty vial is identity by LC-MS, together with content per vial.

Low content: underdosed and overfilled vials

A vial can hold the right peptide in the wrong amount. A content test measures mg of peptide against a reference standard and gives the result as a share of the label. Under Finnrick’s default standard, overfill fails as well as underfill.

Overfill is common in published reports, partly because of how labs report. Peptigrity, which collects certificates, found that about 40% of single-compound certificates that print both the label and the measured amount read above 110% in September 2026. It explains that some labs report the gross mass of the powder, which includes salts and moisture as well as peptide. Weighed against a net label, a gross result can read well above the label for an ordinary vial. In the 2024 JAMA Network Open study, the measured semaglutide content exceeded the label by 29% to 39% in each of the three delivered vials.

Labs also differ on mg. In graymarket’s comparison of one retatrutide batch (3 September 2025), two labs reported about 30 mg and a third reported 26 mg, about 15% lower. That is a single published case, but it shows why a content result needs a stated basis. The guide to underdosed peptides explains how to read the mg result, and content per vial shows what our report states.

Low purity: the right peptide with too much else

Low purity is a quality failure, not a fake. HPLC purity is the share of UV peak area that belongs to the main peptide. The other peaks are other UV-absorbing material, such as shortened or altered versions of the peptide.

In the Vanguard log, these products failed the lab’s 98.9% purity line most often:

Product	Entries below 98.9%
MOTS-c	142 (148 with blends)
Retatrutide	86
NAD+	82
Tesamorelin	78
SS-31	72
BPC-157	71

These are counts of failures, not rates. A product that many people send in can top the list on volume alone. NAD+ is a coenzyme, not a peptide. What a lab checks for MOTS-c, NAD+ and SS-31 is in the guide to less-tested compounds.

The pass line also varies. Vanguard uses 98.9%, while Finnrick’s default standard uses 99.5% and one of its member standards uses 98%. A 99.2% result fails one and passes the other two. Read the number, not only the badge.

The 2024 study shows how far purity can fall. LC-MS found semaglutide in all three delivered vials, but the authors measured purity of 7% to 14% against an advertised 99%.

Peptide contamination: endotoxin and microbes

Contamination is a separate question from identity, content and purity. A vial of the right peptide at full strength can still carry endotoxin or live microbes, and HPLC purity cannot see either.

The published data on this are thin:

• Endotoxin. In the 2024 study, one of three delivered semaglutide vials had 8.95 EU/mg, which the authors called elevated. No viable microorganisms were detected in the three samples.
• Rating standards. Of Finnrick’s five standards, only Accuracy Focus grades endotoxin. It fails a sample above 40 EU per sample, and only when an endotoxin result exists.
• How often it is tested. On Disclosed Labs’ lab table on 4 October 2026, the share of certificates with an endotoxin result ranged from 12% to 89% across the four labs with the most certificates. Where endotoxin is an add-on, that share shows what clients ordered.

So many published results say nothing about endotoxin, either way. Endotoxin needs its own test by LAL, reported in EU per mg or per vial; see endotoxin testing and the guide to endotoxin testing for peptides. Live microbes need the 14-day USP <71> sterility test. A PCR screen or a microbial count is not a sterility test.

How to test peptides for fentanyl

None of the published data sets in this guide reports a fentanyl result, so they cannot say whether or how often it occurs. Some peptide labs sell a screen. Vanguard Laboratory includes fentanyl screening in its Gold panel, and Janoshik lists a “fentanyl presence” add-on (both checked September 2026). A useful report names the method, the compounds it covers and the cut-off or detection limit.

The CDC describes fentanyl test strips for checking drugs such as cocaine, methamphetamine and heroin. It says they show whether fentanyl is present, not how much. They may not detect some analogues, such as carfentanil. A strip result says nothing about the peptide in the vial. HK Analytical does not offer fentanyl screens.

Counterfeit peptides: fakes of licensed medicines

Counterfeits of licensed peptide medicines are a separate market from research peptides, with their own published cases. In June 2024, the WHO warned of falsified Ozempic (semaglutide) found in the regulated supply chain of Brazil and the United Kingdom in October 2023 and the United States in December 2023. The FDA reported seizing thousands of counterfeit Ozempic units in December 2023, with counterfeit needles whose sterility could not be confirmed. It reported further seizures in April and December 2025. The 2024 study met another kind of fake: three of the six online sellers did not deliver and asked for extra payments, said to clear customs.

Checking whether peptides are legit

A seller’s name, reviews or price cannot tell you what is in a vial. Forum reports, including Reddit threads about fake peptides, are single cases. Give weight to a report you can open at the lab’s own site.

A seller’s certificate covers the sample the seller sent. To know about your vial:

1. Test your own vial, or one from the same lot that you picked yourself.
2. Order identity and content first. They catch the two fakes and the fill failures.
3. Add purity, endotoxin and sterility when the question needs them.
4. Test several vials if a batch is shared. A group test splits the cost across buyers.

The guides on how to tell if peptides are real and signs of a fake peptide COA cover the checks in detail.

What HK Analytical does

We run the tests that catch each failure type, in one order and in any mix per sample: identity (LC-MS), blind GLP-1 identification, content per vial, purity (HPLC), endotoxin (LAL), Sterility (USP <71>, 14 days), residual solvents, counter-ion, water content and vial-to-vial consistency. Results within days of receipt. Sterility (USP <71>, 14 days) is the one test that takes 14 days, because the method incubates for 14 days. Every number comes with its raw data. The report carries our photo of the vial as received and states how the client is related to HK Analytical, inside a signed record you can check at hkanalytical.com/verify. The example report shows the layout. Quotes are on request, and group and batch orders are quoted per sample.