Peptide lab comparison tests: what same-batch results from several labs show, and what a fair round needs

What published peptide lab comparisons show

A peptide lab comparison test sends vials from one batch to more than one lab. Five such cases have been published since September 2025, by a Substack writer, a review site and an investigative newsletter. They cover retatrutide, tirzepatide, tesamorelin and epitalon, with two to five labs per batch. None was run as a controlled round. Read together, they show that labs which repeat their own results well can still sit far apart from each other on mg per vial. They cannot show which lab was right.

In our September 2026 market research, we found no published round in which one batch goes to many labs under a protocol set in advance.

The published same-batch cases at a glance

Published case	Product and label	Labs (vials)	mg per vial	Purity	Highest mg above lowest
graymarket, 3 Sep 2025	Retatrutide, one batch	Janoshik (1, plus 9 earlier), Chromate (1), Krause Analytical (2)	About 30; 30; 26 and 26	Not compared	15%
Peptide Critic, 13 May 2026	Retatrutide, one 10 mg kit, plus the seller’s report	Vanguard (2), Analytical Formulations (2), Janoshik (2, plus 3 on the seller’s report), TrustPointe (2), Freedom Diagnostics (2)	10.18 to 12.80	99.74 to 99.93%	26%
Krysia’s Substack, 2 Aug 2026	Tirzepatide, 60 mg	Chromate (2), Accumark (2), BTLabs (1)	62.3 to 75.44	Reported as high at all three	21%
Krysia’s Substack, 13 Aug 2026	Tesamorelin, one batch	Freedom Diagnostics (2), Janoshik (3)	10.055 and 11.057 (means)	99.985% and 97.779% (means)	10%
Same piece	Epitalon, 50 mg	Freedom Diagnostics (2), Chromate (2), Accumark (2)	47.325, 55.525 and 83.64 (means)	99.935%, 95.915% and 99.27%	77%

The last column is our arithmetic. Each row is a single case, and none proves that a lab is wrong.

Labs that sat apart in one case agreed in another. Chromate matched Janoshik on the graymarket batch. A Peptide Critic article of 21 June 2026 set a Krause Analytical result of 59 mg at 99.8% purity beside a Janoshik report of 58.30 to 58.56 mg on three vials of the same seller’s retatrutide, about 1% apart. The article gives no batch number, so it is weaker evidence than a same-batch split.

Within each lab, results were close

The cases hold a second data set that is easy to miss: labs that tested more than one vial.

Case	Lab	Results (mg)	Spread inside the lab
Retatrutide, 10 mg kit	TrustPointe	12.296, 12.328	0.3%
Retatrutide, 10 mg kit	Vanguard	10.18, 10.30	1.2%
Retatrutide, 10 mg kit	Analytical Formulations	10.71, 10.85	1.3%
Retatrutide, 10 mg kit	Freedom Diagnostics	12.47, 12.80	2.6%
Retatrutide, 10 mg kit	Janoshik (seller’s report)	11.03, 11.14, 11.46	3.9%
Tirzepatide, 60 mg	Chromate	70.99, 71.63	0.9%
Tirzepatide, 60 mg	Accumark	73.72, 75.44	2.3%
Retatrutide (graymarket)	Krause Analytical	26, 26	Same value

Every lab that reported two or more vials separately kept them within 4% of each other, and that spread includes any difference in fill. Between labs, the means on the 10 mg kit ran from 10.24 mg at Vanguard to 12.64 mg at Freedom Diagnostics, 23% apart. Each lab repeated its own result well, and the labs sat at different levels. A steady gap like this points to a fixed difference, such as the basis of the mg or the reference standard, more than to random error. In method-validation terms it is a reproducibility question, which our guide to peptide test accuracy separates from repeatability inside one lab. Our report on why labs disagree covers the causes.

What each case can and cannot show

graymarket, September 2025: retatrutide through Finnrick

The author sent four vials of one retatrutide batch to Finnrick and asked for them to go to four different labs. Janoshik and Chromate each tested one vial, and Krause Analytical tested two. Janoshik had already tested nine vials of the batch for the seller and for customers, all at about 30 mg. Janoshik reported about 30 mg, Chromate 30 mg and Krause 26 mg on both vials, which the author called a 15% difference.

The case shows two labs agreeing at one level and a third repeating itself at another. It gives no purity, no stated basis for the mg and one or two vials per lab. Krause states in its questionnaire answers that a routine result comes from one preparation and one injection, with a check-standard acceptance of ±10%, and it offers more on request. One case cannot separate a lab offset from the basis or the vials, and in the June 2026 pairing above, Krause and Janoshik agreed.

Peptide Critic, May 2026: six reports on one retatrutide kit

This is the widest case. Peptide Critic sent two vials each from one 10 mg retatrutide kit to Vanguard, Analytical Formulations, Janoshik, TrustPointe and Freedom Diagnostics. It set the results beside the seller’s original Janoshik report on three vials. The article notes that the labs tested different vials, so the gaps hold variation between vials as well as between methods. All five labs confirmed retatrutide. Every purity result fell between 99.74% and 99.93%, and every mg result sat at 102 to 128% of the label. Peptide Critic’s reading was that the labs agreed the peptide was highly pure and disagreed more on how much was in the vial.

The case shows that a clean sample can give matching purity across methods. It cannot rank the mg results, because the reports were not ordered under one protocol and no reference value exists.

Krysia’s Substack, August 2026: tirzepatide labelled 60 mg

Five vials from different kits of one batch went to three labs. BTLabs reported 62.3 mg. Chromate’s two vials averaged 71.31 mg and Accumark’s 74.58 mg. That is 104%, 119% and 124% of the label. BTLabs pointed to different methods and standards. Janoshik and Vanguard did not test this batch, but they described their practice in the piece. Janoshik standardises its own reference material by CHNS elemental analysis and NMR and reports net peptide. Vanguard buys certified standards from several sources and may use ion chromatography to identify the counter-ion. The author wrote that the results cannot establish whether the true content was near 62 mg, 71 to 75 mg, or varied across the batch.

Krysia’s Substack, August 2026: tesamorelin and epitalon

On tesamorelin, Freedom Diagnostics tested two vials and Janoshik three. Their means were 99.985% and 10.055 mg against 97.779% and 11.057 mg. The author noted that uneven fill could explain part of the mg gap, and that the raw integration data would be needed to settle the purity gap.

On epitalon, three labs each tested two vials. The means were 95%, 111% and 167% of the 50 mg label. Chromate used a HILIC method, and the piece asked whether each method held epitalon long enough to separate it from its impurities. The case shows that purity can differ by method for some peptides. Without the raw data, it cannot show which figure is right.

Why none of these cases can rank labs

• Different vials. Each lab tested its own vials, so fill differences sit inside every gap.
• Different dates. In two cases, some results came from earlier tests ordered by the seller or by customers, so storage and shipping may have differed.
• Few vials, single injections. Most labs tested one or two vials, and several run one injection per sample, so each lab’s own margin is a large part of any gap.
• No reference value. Nobody knew the true content. The majority is not proof, and the lab apart from the others may be the right one.
• No common basis. Labs reported mg on their own terms: net peptide, salt or powder. Our guide to net peptide content explains the difference.
• Which cases get published. Surprising results get written up more often than dull ones (our reading).

What a fair inter-lab round would look like

Analytical chemistry already has a format for this: proficiency testing, a kind of interlaboratory comparison. Eurachem’s 2021 guide describes it. A provider sends sufficiently homogeneous and stable test items to every participant, collects the results and scores each lab against criteria set before the round. A “blind” round goes further: the item looks like a normal customer sample. ISO/IEC 17025 requires accredited labs to take part in proficiency testing or another interlaboratory comparison. For peptides, the August 2026 tirzepatide piece proposed the same idea: “one homogeneous solution to be divided into matched samples” for several labs.

Element	Published cases so far	A fair round
Organiser	A buyer, a writer or a review site, after the results	An organiser that is not a scored lab, with a written protocol before samples go out
Samples	Different vials, sometimes different kits	Random sub-samples of one material, checked for homogeneity and stability first
Timing	Often different dates	Sent to every lab at the same time
Blinding	Not stated	Coded samples; labs do not know the expected value or other labs’ results
Reporting	Each lab’s own basis	Net peptide on one defined basis, the reference material and how its content was assigned, corrections for counter-ion, water and solvents, and an uncertainty
Reference value	None	An assigned value set by a stated route
Scoring	Highest against lowest	A score for every lab against a spread fixed in advance
Publication	Selected cases	Every lab’s result, raw data and score, over several rounds

Setting the reference value

The Eurachem guide lists five routes to an assigned value, following ISO 13528: by formulation, from a certified reference material, from one laboratory, from a consensus of expert laboratories, or from a consensus of the participants. For peptides, a solution made up by weight from a well-characterised reference material would give a value known before any lab reports. A participant consensus is simpler but carries any bias the labs share (our reading on both points).

Scoring each lab

The most common score is the z-score: the lab’s result minus the assigned value, divided by a standard deviation fixed for the round. Eurachem gives the usual reading: 2.0 or less either way is satisfactory, between 2.0 and 3.0 is questionable, and 3.0 or more is unsatisfactory.

A worked example, with figures chosen for illustration: the assigned value is 10.00 mg, and the round allows a standard deviation of 0.50 mg (5%). A lab reporting 10.6 mg scores 1.2. A lab reporting 11.2 mg scores 2.4, questionable. A lab reporting 8.4 mg scores −3.2, unsatisfactory. The guide prefers a spread fixed by fitness for purpose over one taken from the round’s own results, so scores stay comparable from round to round. It also says one unsatisfactory result does not make a lab poor, and that how a lab responds tells more than the result. That is why a fair round repeats.

Choices specific to peptides

• Material. Vials from one batch carry fill variation, so the organiser first measures several vials and reports their spread. One solution split into matched samples removes most of it, if the peptide stays stable in transit.
• What to score. Identity, purity and mg per vial, each on its own. An item whose identity is not given to the labs also tests identity methods.

How to run a smaller comparison yourself

A buyer group can borrow the core of this design:

1. Buy several vials from one batch, number them and assign them to labs at random.
2. Send at least two vials to each lab in the same week. Our guide on how many vials to send covers the count.
3. Withhold the label amount where the lab allows it.
4. Ask every lab for the same things: the basis of the mg, the reference standard, the uncertainty and the chromatograms.
5. Agree before the results arrive that you will publish all of them, and give each lab the chance to comment.
6. Compare the gap between labs with the spread inside each lab and with the stated uncertainties.

A group test can share the cost, and our guide to running a peptide group test covers the money and the vial handling.

What HK Analytical reports for a comparison

Our content per vial result is the mg of the target peptide as free base, measured against a reference standard, with the uncertainty and the % of the label claim. Counter-ion and water content can go in the same order, and vial-to-vial consistency gives each vial’s mg with the mean, range and relative standard deviation. Every number comes with its raw data, including the chromatograms. Each report is a signed record that carries our photo of the vial as received and states how the client is related to HK Analytical. Reports are never edited: a change makes a new version and marks the old one superseded. We do not hold ISO/IEC 17025 accreditation. Our comparison page sets our report format beside five other labs, and the example report shows the layout.