"Best peptide stack" is a shopping query, and the honest answer to it is a research question: has anyone tested the combination? For almost every stack sold, the answer is no. This page applies one criterion to the four combinations people actually search for, and reports where each lands.
The criterion, stated before the results so it can be checked against them:
- Does a study exist that administered the combination, as a combination?
- In what species, and at what scale?
- Did the combination beat its best single component — the only comparison that makes a stack a stack?
- Do the individual components have genuine registered human trials?
No score is given, no vendor or product is rated, and nothing was tested for this page. It is a reading of the indexed literature and of two public trial registries, done on 21 September 2026.
The ranking
| Stack | Combination study? | Species | Beat best single component? | Component trial depth |
|---|---|---|---|---|
| MK-677-based stacks | None | — | Untested | Deepest — 8 registrations, Merck trials, a live Phase 3 |
| BPC-157 + TB-500 ("Wolverine") | One | Rats (n = 32) | No | BPC-157: 4 registrations, none completed in humans for this use · TB-500: none genuine |
| CJC-1295 + ipamorelin | None | — | Untested | CJC-1295: 1, terminated · Ipamorelin: 2 completed, for post-operative ileus |
| GHK-Cu, MOTS-c, SS-31 blends | None | — | Untested | Narrative-review level only |
Read down the third column and the ranking resolves itself: one row has an answer, and the answer is no.
The only combination anyone has tested
In 2026, a group published the first and so far only factorial test of a marketed peptide stack. Thirty-two male Sprague-Dawley rats, each twelve weeks old and around 330 g, underwent standardised Achilles tendon transection and repair and were randomly assigned to four groups of eight: control, BPC-157 at 10 µg/kg/day, TB-500 at 60 µg/kg/day, and BPC-157 plus TB-500. Treatment was intraperitoneal for four weeks, after which tendons were taken for biomechanical testing or for histology, histochemistry and immunohistochemistry (Jt Dis Relat Surg 2026, PMID 42542926).
The design is the right one — separate arms for each component and for the pair is the only way a combination can be shown to add anything. The result is the part nobody quotes:
| Measure | BPC-157 | TB-500 | Combination |
|---|---|---|---|
| Maximum load to failure vs control | Higher, not significant | Higher, significant (p < 0.05) | Not reported as significant |
| Bonar score (lower is better) | — | Significant (p = 0.016) | — |
| Movin score (lower is better) | Numerically lower, not significant | Significant (p = 0.017) | Significant (p = 0.040) |
On the biomechanical endpoint — the one that asks whether the tendon is stronger — only TB-500 alone reached significance. On the histology scores, TB-500 alone was significant on both systems, and the combination on one, at a weaker p-value. The combination never separated from its stronger component on any measure reported.
That is a rat study with eight animals per group, and it should not be over-read in either direction. What can be said precisely is this: the only experiment ever run on the most-sold peptide stack tested it properly, and it did not find that the stack beat half of the stack. Everything else written about that combination is reasoning about two mechanisms.
This site's separate page on what exists in humans for that pairing, and what does not covers the human side.
What the registries say about the components
A stack can only be as good as its parts, and the parts are thinner than their marketing.
TB-500 has no genuine registered trial. ClinicalTrials.gov returns exactly one record, NCT07487363, a Phase 1/2 study of TB-500 against placebo in stable atherosclerotic cardiovascular disease. Its brief summary opens: "This fictional study is an example of a ClinicalTrials.gov-style record." Nothing in the record's status, phase or enrolment fields reveals that — the disclaimer is in prose, and a search that counts registrations without reading them reports one trial where there are none.
BPC-157 has four registrations and not one completed human efficacy result. NCT02637284 (PharmaCotherapia) has an unknown status. NCT07437547, a Phase 2 hamstring-strain trial planning 120 participants, is recruiting and carries no such disclaimer. NCT07752381 is completed — 40 physically active adults taking BPC-157 gummies for eight weeks, sponsored by the product's maker. And NCT07803250, a 30-participant pilot at the University of Arkansas in patients undergoing arthroscopic rotator cuff repair, is not yet recruiting; its own summary states that BPC-157 "has not yet been studied in formal human clinical" trials.
CJC-1295 has one registration and it was terminated — a multicentre Phase 2 study in HIV patients with visceral obesity, sponsored by ConjuChem, planning 120 patients at low dose, high dose or placebo.
Ipamorelin has three registrations and two of them are about bowels. Both completed Helsinn trials tested whether post-operative ipamorelin speeds recovery of gastrointestinal function after small or large bowel resection, the larger enrolling 320 patients. The third registration, returned by the same keyword search, is an observational study of ibogaine, 5-MeO-DMT and magnetic resonance therapy in Special Operations Forces veterans with PTSD and traumatic brain injury; ipamorelin appears there under an intervention labelled physiological supplementation. A registry search returns records that mention a compound, not trials of it — counting the three as ipamorelin trials would overstate the field by half.
Why no human study exists for the CJC-1295 and ipamorelin pairing, and what the ten papers that mention both actually are, is covered on this site's page about that stack.
The exception, and why it is not the answer either
MK-677 (ibutamoren) is the only stack component with a real clinical file. Eight registrations, including a Merck randomised trial in post-hip-fracture sarcopenia, a Merck trial in Alzheimer disease, an MGH study in non-alcoholic fatty liver disease, University of Virginia studies of lean body mass and IGF-1, and a live Phase 3 programme in children with growth hormone deficiency.
It is also, for exactly that reason, the compound whose ceiling is best known. The trials that established MK-677 raises IGF-1 and adds lean mass are the same trials that established how much: roughly a kilogram, without an accompanying strength gain. What the human trials measured sets out the numbers, and the class as a whole shows the same pattern across every secretagogue that has been tested.
So the component with the best evidence is the one whose evidence sets the lowest ceiling. That is not a paradox; it is what happens when a class is actually measured. The stacks with no ceiling reported are not the ones that beat it — they are the ones nobody has measured.
Why stacks are untested, structurally
Testing a combination properly needs four arms, which means roughly twice the participants for the same statistical power, which means roughly twice the cost. Nobody funds that for a compound they cannot sell as a drug. The one group that did it used rats.
The consequence is that every claim of synergy in this market is a claim about mechanisms rather than outcomes. Mechanistic reasoning is how good trials get designed, and it is not a substitute for running them — a distinction that also runs through the comparison with steroids, where only one side was ever tested and through the SARM class, which ran two Phase III trials for muscle and still failed.
One thing every stack has in common
All of it is prohibited in sport, at all times, and under three different sections of WADA's 2026 Prohibited List: growth hormone releasing factors and secretagogues under S2.2.4, thymosin-β4 and its derivatives under S2.3, and BPC-157 under S0 as a substance no health authority has approved. Those sections carry different sanction classes, which is why the section-by-section register is worth reading before "is it banned?" gets answered from memory.