CJC-1295 fails purity more often than any other widely tested peptide in the lab reports we track: 6.25% below 90%, vs 0.21% for tirzepatide

The short answer

CJC-1295 has failed purity more often than any other peptide with at least 300 lab reports in TitrateLab’s records. We compared all 36 product labels at that volume, including blends and a few non-peptides such as NAD+. Measured on October 1, 2026, 60 of 960 CJC-1295 reports with a purity result (6.25%) were below 90%. Under the same rule, tirzepatide was 24 of 11,272 (0.21%) and retatrutide 54 of 17,360 (0.31%). Those two are the most-reported peptides in our records. The next-highest peptide with 300 or more reports was ARA-290, at 8 of 479 (1.67%).

The finding has a hard boundary. All 60 failures are Finnrick records, and 54 were tested in 2025. The 663 CJC-1295 reports from other sources include none below 90%. The lab alone does not explain that split. Janoshik and Freedom Diagnostics issued 653 of those 663 certificates, and on CJC-1295 samples Finnrick sent them in 2025 they reported 11 of 38 below 90%.

What the record supports:

• Within Finnrick’s data, CJC-1295 failed far more often than the GLP-1s: 60 of 297 (20.20%), against 0.80% for tirzepatide and 0.90% for retatrutide.
• The gap holds both for vials Finnrick bought from vendors and for vials members of the public mailed in.
• Failures came from seven of the eight labs Finnrick used for CJC-1295, so no single lab accounts for them.

What it does not establish:

• a current, market-wide failure rate for CJC-1295;
• why these samples failed;
• anything about a batch that was not tested.

TitrateLab does not run these tests. Every figure comes from a third-party lab report that we collect, structure and link.

How CJC-1295 compares

The size of the gap depends on how it is counted, so here are four counts from the same records:

The last row needs explaining. Finnrick marks some samples “Identity Fail - not the labeled substance”, and our records store their purity as 0%. Our headline counts only reports with a purity figure above 0, so records at 0% are left out: 36 for CJC-1295 (all Finnrick identity failures), 133 for tirzepatide (128 from Finnrick) and 289 for retatrutide (285 from Finnrick). Counting them as failures raises every rate and shrinks the CJC-1295 multiple to about 7 and 5 times.

A stricter 95% bar gives the same ordering: CJC-1295 96 of 960 (10.00%), tirzepatide 38 of 11,272 (0.34%), retatrutide 108 of 17,360 (0.62%).

Where the 60 failures come from

By source

Freedom Diagnostics certificates print the client who ordered the test, and our records attribute the vendor from that name: 453 of the 562 carry a named vendor and 109 do not. A business client usually means the seller commissioned the test, but some clients are buyers or group organizers. Janoshik certificates print a client too. In our records 74 of the 91 have a client name, one of them “Anonymous”, and the other 17 have none recorded. We attribute 66 of the 74 to a vendor, and in most of those the vendor is the business named as client, which usually means the seller commissioned the test.

By lab

Finnrick’s record pages give the code and name of the lab that ran each test, and its Partner Testing Labs page lists the labs with their test counts and dates. On October 1, 2026 we checked the lab name on Finnrick’s page for all 11 Janoshik and Freedom Diagnostics failures and for at least one record under every other lab code.

Janoshik and Freedom Diagnostics also issue certificates to their own clients, and those make up most CJC-1295 reports outside Finnrick. The comparison inside each lab:

• Janoshik reported 8 of the 26 CJC-1295 samples Finnrick sent it below 90%, all tested April to August 2025. None of its 91 client certificates was below 90%, including 0 of 17 tested April to September 2025.
• Freedom Diagnostics reported 3 of the 12 samples Finnrick sent it below 90%, all tested April to August 2025. None of its 562 client certificates was below 90%, including 0 of 30 tested April to September 2025.

All 11 of those failures were vials Finnrick bought. The same labs reported failures on Finnrick’s vials and none on their clients’ vials, so the absence of failures outside Finnrick is not explained by those labs never reporting one. The vendors differ between the two groups, and 527 of Freedom Diagnostics’ 562 client certificates were tested after September 2025, when Finnrick’s own CJC-1295 failure rate was lower: 15 of 118 after September 2025, against 45 of 179 before.

By who sent the sample

Finnrick labels every record by who supplied the sample. Its record pages use these words for the three labels that cover every CJC-1295 record:

• Finnrick: “This sample was ordered by Finnrick directly from the vendor, anonymously or not.”
• Public: “A member of the public mailed in this sample for testing.”
• Vendor: “The vendor submitted this sample for testing.”

A fourth label, “Pulse”, covers 6 tirzepatide and 10 retatrutide records, none below 90%, and no CJC-1295 records. It is left out of the table, so the GLP-1 rows sum to 2,603 of 2,609 and 4,090 of 4,100 Finnrick records.

Vials Finnrick bought are the closest these records come to an ordinary purchase from the vendor, and that is where CJC-1295 failed most often. Public submissions carry a different bias: people may send in vials that already look wrong. CJC-1295 still failed about 11 to 13 times as often as the GLP-1s within that group.

By test date

54 of the 60 failures were tested in 2025, 44 of them between January and June. By test year, 54 of 350 reports dated 2025 failed (15.43%) and 6 of 581 dated 2026 failed (1.03%).

That drop mostly reflects a change in which reports exist. Finnrick-bought CJC-1295 reports fell from 160 tested in 2025 (45 failed) to 2 dated 2026, and those two appear to be one test recorded twice (see the method below). Meanwhile 504 of the 581 reports dated 2026 come from other sources, and none of them failed. Had each source kept its 2025 failure rate, the 2026 mix alone would give about 3.3%.

The rate also fell within Finnrick’s records, from 54 of 218 (24.77%) in 2025 to 6 of 77 (7.79%) in 2026. Among public mail-ins it fell from 9 of 54 (16.67%) to 4 of 65 (6.15%). The 2026 Finnrick rate is still about six times the GLP-1s from the same source that year: 18 of 1,454 for tirzepatide (1.24%) and 31 of 2,618 for retatrutide (1.18%). Over the same two years, Finnrick’s tirzepatide and retatrutide rates rose, from 0.26% and 0.41% in 2025. Our community-testing reanalysis found the same shift in CJC-1295’s source mix.

The ten lowest-purity CJC-1295 reports

Each row is one tested sample on one date, not a vendor’s whole catalog.

The Mai Peptide report entered our records on September 30, 2026. It replaces Loti Labs’ 60.95% report (/research/coa/1428) in this list.

Purity and amount are separate results. Finnrick reports how pure a sample was and, separately, how much material it measured against the label. On Finnrick’s own pages, checked October 1, 2026, the Swiss Chems sample shows 1.692% purity with a fill 35% under its 2 mg label, and the Mai Peptide sample shows 17.82% purity with a fill 3% over its 5 mg label. Both pages show an identity pass. They do not explain how one sample produced both figures, so we do not estimate how much CJC-1295 either vial held.

Vendors with three or more failing reports

This table lists every vendor in our records with three or more CJC-1295 reports below 90%, using the headline rule. Each row links that vendor’s lowest-purity report.

These 11 vendors account for 40 of the 60 failures. 24 vendors have at least one. Every failing report in the table is a Finnrick record.

Eight of the 11 also have at least one CJC-1295 report at or above 90%. Three do not: Loti Labs (0 of 5), Planet Peptide (0 of 4) and Swiss Chems (0 of 3). All 12 of those reports are vials Finnrick bought, tested between February and May 2025, and we hold no later single-label CJC-1295 report for any of the three. This covers single-label CJC-1295 only. Planet Peptide also has eight CJC-1295 + Ipamorelin blend certificates from Freedom Diagnostics, tested September 2025 to July 2026, all above 99% purity.

The table does not count identity failures. Finnrick’s listing on October 29, 2026 still showed four for vendors in it: Suaway Lab Research 2 (23645 and 23646, both tested April 1, 2025) and Wuhan Wansheng Biotechnology 2 (23650 and 23649, tested January 26 and March 6, 2026). On October 1 we also listed one for Liberty Peptides. On October 15, 2026 Finnrick changed that result to an identity pass, though it still grades the sample FAIL overall, for the reason given under explanation 4, which molecule, below.

Small counts amplify. A vendor with three reports that all failed has a 100% documented failure rate, but three is three, and those three were tested in 2025. Read the certificates rather than the rate.

What could explain the gap

The observation is narrow: CJC-1295 samples tested through Finnrick failed purity far more often than tirzepatide or retatrutide samples tested through Finnrick, among vials Finnrick bought and among vials the public sent in. The records do not say why. Four explanations remain open. The lab comparison above makes a lab-only explanation unlikely; the data cannot rank the other three.

1. Sample selection. Who chose the vial differs by source. Finnrick bought 164 of its 297 CJC-1295 samples from vendors, and members of the public mailed in most of the rest. Freedom Diagnostics certificates are ordered by a named client, usually a seller. Janoshik and Freedom Diagnostics reported failures on Finnrick’s vials and none on their clients’ vials tested over the same period, which fits this explanation. Our domestic-sellers analysis found the same direction for US sellers across many compounds: seller-commissioned certificates failed far less often than vials Finnrick bought or customers mailed in.
2. Manufacturing or storage. Some CJC-1295 sold by these vendors in early 2025 may have been poorly made or degraded. Failures cluster by vendor and month, which fits this explanation, but no certificate states a cause. The vendors behind Finnrick’s samples and the labs’ client certificates differ, so this cannot be separated from sample selection.
3. Laboratory and method. Failures came from seven different labs, and the two labs with many client certificates in our records reported failures only on Finnrick’s samples. One lab’s method does not explain the pattern. Labs could still differ in how they measure CJC-1295 purity, and these records cannot test that.
4. Which molecule. CJC-1295 is sold with and without DAC. Our records keep the bare label as printed. Finnrick’s pages show some of these reports as “CJC-1295 (No DAC)”, including the Swiss Chems and Mai Peptide reports above, and others as “CJC-1295 (DAC)”, including Astro Peptides reports 1399 and 1400. If a sample was tested as the other form, a low purity or failed identity result could partly reflect that mismatch. Finnrick has documented at least one such case. On October 15, 2026 it changed the identity result for a Liberty Peptides sample first dated April 23, 2026 from a fail against CJC-1295 (no DAC) to a pass against CJC-1295 (DAC), and reported 90.8% purity and 8.73 mg measured against a 10 mg label. Finnrick’s page still shows the overall result as “FAIL”. Its correction says “the CJC variant was inconsistently communicated to the lab by Finnrick, and the lab ran its initial analysis against the incorrect variant”. It adds that the vendor raised a formal objection, Finnrick opened an investigation, and the lab then issued an updated analysis based on the original measurements. Finnrick now dates the record October 15, 2026; our copy is 119828. The original failure was one of the identity failures counted on October 1, not one of the 60 purity failures. On October 29, 2026, Finnrick’s listing still showed the results we hold for the other 36 CJC-1295 identity failures counted on October 1 and for 295 of the 297 Finnrick purity results; the other two are the earlier addresses described under possible duplicates. A matching listing does not show which form the lab tested against, so the records still cannot show how often a mismatch happened.

What a certificate can and cannot tell a buyer

A failing report describes one sample on one date. Eight of the 11 vendors in the table also have passing CJC-1295 reports, and most failures are from 2025. A vendor’s past results do not tell you which result your vial would get.

The closest evidence to a specific vial is a report for that batch from a sample the seller did not choose. Most passing CJC-1295 reports in our records (562 of 900) are Freedom Diagnostics certificates ordered by a named client, usually the seller. That does not make them wrong, but they answer a narrower question: how the vial the client sent tested. Our guide on how to read and check a COA covers what to compare, and the COA analyzer structures a certificate you already hold. The CJC-1295 research hub shows the current aggregate counts, and each certificate cited above links to its record.

Method and evidence boundary

• Records. Published records in TitrateLab’s COA database whose peptide label is exactly “CJC-1295”, with a purity figure above 0, added to our records before October 1, 2026 (00:00 UTC). Measured October 1, 2026. Tirzepatide and retatrutide use the same rule with their exact labels. New reports arrive daily, so later counts will differ.
• Failure. Purity below 90%. We use 90% because it is lenient and unambiguous. The 95% figures are given above.
• 0% purity records. Records with purity stored as 0, mostly Finnrick identity failures, are counted separately, as shown above.
• Blend flags. 52 of the 960 records carry the bare label but are flagged internally as blends. Excluding them gives 60 of 908 (6.61%).
• Other CJC-1295 labels. Records labeled “CJC-1295 w/ DAC”, “CJC-1295 w/DAC”, “CJC-1295 no DAC” or “CJC-1295 (mod GRF 1-29)” are not in the headline. Adding them gives 65 of 1,357 (4.79%). Combination vials such as “CJC-1295 + Ipamorelin”, and rarer spellings with fewer than five records each, are excluded.
• Vendors. Names are as attributed to each report. 121 of the 960 records have no identified vendor: 13 are blank and 108 carry a placeholder (“Unknown vendor (review)” or “(unknown)”). The rest carry 392 distinct names, a few of which are client labels rather than clear vendor names, such as “Anonymous”.
• Dates. Test dates run from November 10, 2019 to September 27, 2026. 28 records are dated before 2025, and one has no test date.
• Possible duplicates. Some records repeat another record’s result, among passing and failing reports alike. Two Finnrick pairs have addresses that differ only by date, and on October 1, 2026 the earlier address of each redirected to the later one: failing Hydro Research certificates 108630 and 109093 (89.98%, now dated September 20, 2026) and passing Hxnet certificates 106990 and 113164 (99.6%). Other groups share the same source, purity, measured amount, batch label and test date: failing Astro Peptides certificates 1399 and 1400 (88.48%, April 25, 2025), two Janoshik pairs, and 31 Freedom Diagnostics groups in which 82 certificates repeat one result under different client names. Treating every such group as one test gives 58 of 904 (6.42%). We count records as published.
• Not a random sample. Every record is a test that someone chose to order, submit or publish. The figures describe those reports, not the whole market.
• Labs. Our Finnrick records store the lab code shown on each Finnrick page. We matched codes to lab names on Finnrick’s own record pages, as described above. On October 29, 2026, the codes for Freedom Diagnostics, Janoshik and MZ Biolabs matched the test counts and date ranges on Finnrick’s Partner Testing Labs page exactly, as TrustPointe’s code did on October 1. On October 29, TrustPointe’s dates also matched, but Finnrick listed 313 of its tests, two fewer than the 315 in our records. The two are Liberty Peptides retatrutide tests, both above 98% purity, that Finnrick stopped listing between October 26 and 27; Finnrick has not said why.
• Sources. Results come from Finnrick, Freedom Diagnostics, Janoshik and other public certificates. Each certificate link opens TitrateLab’s record of that report.

Corrections and updates

October 29, 2026.

• Liberty Peptides identity failure. On October 1 we listed a Finnrick identity failure for Liberty Peptides among five for vendors in the table. Finnrick changed it on October 15, 2026: because Finnrick had communicated the variant to the lab inconsistently, the lab first analyzed the sample against the wrong CJC-1295 form. The corrected record shows an identity pass with 90.8% purity, and Finnrick still grades it FAIL overall (119828). It entered our records on October 16, after the October 1 cutoff, so the tables do not count it, and Liberty’s row stays at 3 failing reports of 4. Liberty’s other April 23, 2026 report, the purity failure 12543 (83.8%), is a different sample and batch, and Finnrick still lists it. We no longer publish the original failure, and it is removed from the counts above. CJC-1295 records stored at 0% purity fell from 37 to 36, and the rate with them counted from 97 of 997 (9.73%) to 96 of 996 (9.64%).
• Method details. The Janoshik client figure now states how it is counted: 74 of 91 certificates carry a client in our records, and 66 of those are attributed to a vendor. The check of lab codes against Finnrick’s labs page is now dated. On October 29 three codes still matched exactly. Finnrick’s count for TrustPointe was two lower than ours, because it stopped listing two Liberty Peptides retatrutide tests.
• Re-run. We re-ran our queries on October 29 with the same October 1 cutoff. Apart from the Liberty Peptides record, the CJC-1295 figures were unchanged except vendor attribution. Three Freedom Diagnostics certificates, all above 99%, now show no vendor while their attribution is reviewed. So 16 records have a blank vendor, not 13, and 124 have no identified vendor, not 121; 391 distinct vendor names appear, not 392; and 450 of the 562 Freedom Diagnostics certificates carry a named vendor and 112 do not, not 453 and 109. Tirzepatide and retatrutide counts moved by at most six records, which changed none of their rates or means by more than 0.01 points. Apart from the method details above, we changed only the figures that depended on the corrected record; every other figure remains as measured on October 1.

October 1, 2026.

• Liberty Peptides. The July 31 version said Liberty Peptides had no CJC-1295 report above 90%. That was wrong when it was published. Freedom Diagnostics certificate 39094 (99.11%, tested February 27, 2026) entered our records on July 4, 2026. Liberty’s record under the headline rule is 3 failing reports of 4. This entry also listed one Finnrick identity failure, which Finnrick changed to an identity pass on October 15, 2026 (see October 29). Liberty also has three Freedom Diagnostics certificates labeled as the DAC form (“CJC-1295 w/ DAC” or “CJC-1295 w/DAC”), all above 99% (61488, 39022, 39095).
• Vendors with “no failing batch documented”. We removed this list of five vendors. It counted only reports with a purity figure, so it left out Finnrick identity failures for two of the vendors it named, and it linked no certificates.
• Title. The old title called CJC-1295 “the dirtiest peptide we test”. TitrateLab does not test peptides; we collect third-party lab reports.
• Removed statements. The chemistry and market-structure section, which cited no source. The claim that the bare “CJC-1295” label means the no-DAC form; Finnrick shows both forms under it. The claim that the failures are “a synthesis problem, not a testing-noise problem”. The claim that the query file was published alongside the article; it was not served publicly. The line that tirzepatide buyers “can be casual”.
• Figures. The headline moved from 53 of 683 (7.76%) on July 31 to 60 of 960 (6.25%). The first version, in June 2026, reported 51 of 269 (18.96%).

July 31, 2026. Suaway Lab Research was removed from a list of vendors with no CJC-1295 report above 90%. Its fourth report, 55909 (90.44%, tested May 6, 2026), entered our records on July 4, 2026.

Think a figure or attribution here is wrong? Write to [email protected] and include the certificate number.

Open original ×