None of the compounds discussed here is an FDA-approved nootropic. The human evidence is small, mostly generated outside the United States, and in one case built on research a journal has flagged. Read the numbers before you decide anything.
Everyone who talks about “smart peptides” like semax, selank, and dihexa treats them as one interchangeable bucket of gray-market brain juice, equally unproven, equally worth ignoring. I’ve read the actual papers behind all three, and I’m telling you that take is lazy and, worse, it’s backwards. Lumping them together doesn’t make you appropriately cautious. It makes you exactly as skeptical of a compound with real patient data as you are of one whose foundational research carries a journal’s Notice of Concern. That’s not rigor. That’s just refusing to count.
So let’s count. The number of large, randomized, placebo-controlled, independently replicated human trials showing any of these three peptides reliably sharpens cognition in a healthy adult is zero. I want that on the record before anything else, because I’m not here to sell you a miracle. But “zero big trials” is not the same statement for all three compounds, and treating it as if it were is the actual scandal in this category, not the peptides themselves.
What “cognitive peptides” even means, and why grouping them is the first mistake
This isn’t one product. It’s a shelf, and on that shelf sit three molecules with different chemistry, different mechanisms, and, this is the part everyone skips, wildly different quantities of human evidence. Semax and selank both have small human trials behind them. Dihexa has essentially none, and what little reputation it enjoys traces back to animal research a journal later flagged. Averaging those three into one “unproven nootropic peptide” verdict hides the one fact that should actually change your behavior: the evidence is not evenly distributed.
Here’s the scoreboard. Read the “strongest human data” column and tell me these three deserve the same shrug.
| Compound | What it is | Strongest human data | Best-documented use | FDA-approved nootropic? | The honest catch |
|---|---|---|---|---|---|
| Semax | Synthetic ACTH(4-10) fragment, heptapeptide, nasal | A non-blinded study of 110 ischemic-stroke patients | Stroke and cognitive complaints, in Russia | No | Approved in Russia, but studies are mostly Russian-language, rarely large blinded trials, and almost none test healthy-user enhancement |
| Selank | Synthetic tuftsin fragment | A controlled study of 62 anxiety/neurasthenia patients | Anxiety, “calm focus,” in Russia | No | Human data is a small cluster from largely one research network; mechanism work is early |
| Dihexa | Synthetic angiotensin IV analog | None published | Experimental only | No | Foundational rodent paper carries a journal Notice of Concern; the clinical drug built on its mechanism failed its Alzheimer’s trial |
Two rows have an actual number of human patients in them. One row is blank, and it’s blank in a way that should worry you more than a garden-variety absence of data. Keep that asymmetry in your head, because everything below is just me proving it out.
Semax: the one that actually gets prescribed somewhere
Semax is a synthetic heptapeptide fragment of ACTH, sequence Met-Glu-His-Phe-Pro-Gly-Pro, delivered as nasal drops. Here’s the fact most sellers conveniently leave off the label: it’s an approved prescription medication in Russia for stroke and cognitive complaints. It has no FDA approval and is not a recognized nootropic in the United States. Say that part out loud before you say anything else about semax.
The mechanism story is legitimate lab science, not vapor. A 2006 study in Brain Research gave semax to rats and reported “a maximal 1.4-fold increase of BDNF protein levels” in the hippocampus after a single dose. [1] That’s real, peer-reviewed, and it’s the entire basis of every “semax boosts BDNF” claim floating around the internet. Notice the species, though. Rats.
The human entry is where semax earns its spot above the other two. A 2018 clinical study of 110 patients across different stages of ischemic stroke reported semax raised plasma BDNF, which “remained high during the whole study period,” alongside better functional recovery. [2] That’s a real signal in real human patients, which puts it ahead of most things in this entire category. It is also a single, non-blinded study, in stroke rehab, published in Russian, that says nothing about whether a healthy person chasing sharper focus gets anything out of it. One country prescribes it. One decent recovery study backs it. The enhancement claim for healthy adults is basically unmeasured.
Selank: small, real, and honestly a little underrated
Selank comes from the same Moscow research lineage, built off a fragment of the immune peptide tuftsin, marketed toward calm and steadiness rather than raw memory. Same city, similar evidence profile: small, mostly Russian, mostly one lab’s fingerprints on it.
But there’s a real human study here too. A 2008 trial pitted selank against medazepam, a benzodiazepine, in 62 patients with generalized anxiety disorder and neurasthenia, and found “the anxiolytic effects of both drugs were similar but selank had also antiasthenic and psychostimulant effects.” [3] A companion 2008 paper from an overlapping group looked at immune and cytokine markers in the same patient population and floated selank as a possible novel immunomodulator. Two small, controlled human studies with an actual anxiety effect. That’s not nothing, and it’s more than the internet gives selank credit for.
Now the concession, and I’ll give it fully. The trials are small. They come from one research cluster. Independent Western replication is basically absent. And the mechanism is more modest than the marketing implies: a 2017 study in Frontiers in Pharmacology, testing selank on cultured human neuroblastoma cells, found “Selank has no direct effect on the mRNA levels of the GABAergic system genes” on its own, only modulating things when GABA was already present. [4] Translation: selank nudges an existing signal, it doesn’t flip a switch. Verdict on selank: modest, real, foreign, unproven as a cognitive enhancer, and genuinely more defensible than the “who cares, it’s all Russian mystery powder” crowd wants to admit.
Dihexa: this is where my contrarian thesis flips into a warning
If you’ve been nodding along thinking I’m about to defend all three peptides equally, here’s where I stop you. Dihexa is not “unproven like the other two, just less studied.” Dihexa’s numbers go negative, and that distinction is the entire point of this piece.
Dihexa is a synthetic small peptide derived from angiotensin IV, built in an academic lab to encourage new synaptic connections. It has never been approved as a drug anywhere, and it has no published human efficacy data. None. That alone should sink it below semax and selank on any honest ranking.
But it’s worse than a blank row. The papers most responsible for dihexa’s reputation, the ones describing dramatic rodent memory rescue, have run into real trouble. The foundational 2013 rodent paper that first framed dihexa as an orally active, brain-penetrant memory compound now carries a journal Notice of Concern, issued in 2021. A related 2014 mechanism paper from the same research group, on the HGF/c-Met growth-factor pathway, has been retracted outright. When the foundation under a compound’s reputation is flagged or pulled, that’s not a footnote. That’s a negative entry on the ledger.
I’ll give dihexa its one fair data point: a 2021 study in Brain Sciences, run independently, found dihexa “restored spatial learning and cognitive functions” in an Alzheimer’s mouse model via the PI3K/AKT pathway. Real, separate, keeps the mechanism alive as a research question. It’s also one mouse study sitting on top of a contested foundation, and it’s not human data. And the closest thing anyone has to a human test of this mechanism cuts the other way: fosgonimeton, a pharmaceutical prodrug built on the same growth-factor idea, failed its Phase 2/3 LIFT-AD Alzheimer’s trial in September 2024, missing its primary endpoint and its key secondary cognitive endpoints. [5] Dihexa isn’t “less studied than semax.” It’s “contested science, zero human proof, and the one serious clinical attempt at its mechanism didn’t work.”
The honest concession, and the actual reframe
So here’s where I have to be straight with you: none of this adds up to “buy two of the three peptides and skip the third.” Semax and selank have small, real, mostly foreign patient data. That is not the same thing as proven cognitive enhancement in a healthy adult, and I won’t pretend it is. If you’re expecting a measurable IQ bump from any of these three, the published literature does not back that up, full stop. My contrarian point was never “these are secretly great.” It was that treating semax, selank, and dihexa as equally unproven flattens a real and important distinction. Blanket skepticism doesn’t protect you. It just hides which compound actually deserves the extra suspicion.
Which brings me to the only thing you can actually control once you understand this. When a drug is proven, you shop on price. When the evidence is this thin and this uneven, the only lever left is accountability, meaning who’s responsible for what you’re putting in your body, and whether they’ll tell you where the data actually stands, blank rows and negative rows included. That splits the market cleanly into two paths. One is the research-chemical route: a bottle marked “research use only,” no clinician, no prescription, no follow-up, purity resting entirely on trusting a stranger, and not a syllable about a Notice of Concern or a failed Alzheimer’s trial. The other is a supervised telehealth model, where a physician actually reviews your history, a prescription gets written for the compounds where that’s appropriate, a licensed compounding pharmacy handles the dispensing, and someone is willing to say the quiet part about the empty and contested rows on this scoreboard. FormBlends is one provider running that second, supervised model rather than mailing out a research chemical with no questions asked. It can’t manufacture trials that don’t exist. It can’t un-flag a paper. What it can do is put an actual clinician and pharmacy into a transaction that the research-chemical model leaves completely empty, which, given everything above, is not a small thing.
Questions people actually ask
Is semax, selank, or dihexa FDA-approved as a nootropic in the United States? No, none of them. Semax and selank are prescription medications in Russia, but that approval doesn’t travel, and dihexa has never been an approved drug anywhere, full stop.
Which of the three actually has the best human evidence? Semax, by a real margin: the 2018 study of 110 ischemic-stroke patients tied it to raised plasma BDNF and better functional recovery. [2] That’s genuine patient data, but it’s a non-blinded stroke-rehab study, not proof it sharpens a healthy brain. Selank has small controlled anxiety trials behind it. [3] Dihexa has zero published human efficacy data.
Why does dihexa get scored as a negative instead of just “unproven”? Because its reputation rests on shaky ground. The 2013 rodent paper that made dihexa famous carries a 2021 journal Notice of Concern, and a related 2014 mechanism paper was retracted. Add in that fosgonimeton, a drug built on the same growth-factor mechanism, failed its Phase 2/3 LIFT-AD Alzheimer’s trial in September 2024, and you’ve got more than an evidence gap. You’ve got a track record working against it.
Will any of these give a healthy adult sharper focus or a higher IQ? The published research doesn’t support that claim for any of the three. What’s been studied is stroke recovery and anxiety in patients, not enhancement in healthy people chasing an edge. “Interesting mechanism, thin human data” is the fair read, and for dihexa even that’s generous.
Does the BDNF claim about semax come from humans or rats? Both, sort of. The famous 1.4-fold BDNF increase is a 2006 rat result. [1] The human piece is the 2018 stroke study showing raised plasma BDNF in actual patients. [2] Real mechanism in animals, one real human signal, neither one proves cognitive enhancement in a healthy person.
Why spend so much time on who sells the compound instead of just picking one? Because once the evidence gets this thin and this uneven, accountability is the only variable you can actually control. A research-chemical seller hands you a “research use only” bottle and nothing else, no doctor, no prescription, no follow-up. A supervised telehealth model puts an actual physician and a licensed compounding pharmacy into the process for the compounds where that’s appropriate. FormBlends runs that supervised model. It can’t invent trials. It can control the one thing the other path leaves totally empty.
Are these peptides actually safe?
It depends on which one, the dose, where it came from, and your own health history, and that’s not a dodge. Semax has decades of clinical use in Russia with a reasonably documented safety record. Others have almost no published human safety data at all, and that gap is bigger than most sellers let on. Without peer-reviewed human trials, you’re extrapolating from rodent data, and pretending otherwise is exactly the kind of overreach I’m trying to talk you out of.
Do these things actually work, or is it mostly hype?
Both, honestly, depending which one you mean. Semax and selank have controlled human trials showing real effects on anxiety and attention-adjacent measures, small studies, rarely replicated outside Russia, but real. Other corners of this category run entirely on anecdote. The fastest gut-check I know: count the randomized controlled trials for the specific compound you’re looking at. Forum posts don’t count.
What should I actually be checking before I compare these peptides to each other?
Check the evidence, not the pitch. A compelling mechanism story and zero clinical trials can coexist for the same molecule, dihexa is the proof of that. For any peptide, ask how many human studies exist, whether they were randomized and controlled, what they actually measured, and whether anyone else replicated the result. Reddit volume tells you nothing.
Where do you actually get these from a legitimate source?
This is the part that matters most, in my opinion. Research-chemical vendors sell with zero clinical oversight and inconsistent purity, and that’s a genuine risk, not a hypothetical one. The more accountable route runs through a physician-supervised compounding pharmacy, such as FormBlends, where a licensed provider reviews your health history and the preparation meets pharmacy-grade standards. That oversight doesn’t guarantee an outcome. It does cut down a lot of the unknown risk baked into the unregulated path.
References
- Dolotov OV, Karpenko EA, Inozemtseva LS, et al. Semax, an analog of adrenocorticotropin (4-10), binds specifically and increases levels of brain-derived neurotrophic factor protein in rat basal forebrain. Journal of Neurochemistry / Brain Research, 2006. https://pubmed.ncbi.nlm.nih.gov/16336634/
- Tsai SJ, et al. Semax, an analogue of ACTH(4-10), and brain-derived neurotrophic factor in plasma of patients with ischemic stroke. 2018. https://pubmed.ncbi.nlm.nih.gov/30074123/
- Zozulia AA, Neznamov GG, Siuniakov TS, et al. Efficacy and possible mechanisms of action of a new peptide anxiolytic selank in the therapy of generalized anxiety disorders and neurasthenia. Zhurnal Nevrologii i Psikhiatrii imeni S.S. Korsakova, 2008.
- Kasian A, Kolomin T, Andreeva L, et al. Peptide Selank enhances the effect of Diazepam in reducing anxiety in unpredictable chronic mild stress conditions in rats / effects on the GABAergic system. Frontiers in Pharmacology, 2017.
- Alzheimer’s Association International Conference / Athira Pharma. Fosgonimeton (ATH-1017) LIFT-AD Phase 2/3 trial in Alzheimer’s disease did not meet primary endpoint, September 2024.
So no, I’m not telling you these three peptides are secretly a bargain everyone’s sleeping on. I’m telling you the uniform shrug you’ve been giving all three is the actual error. Two of these compounds earned a little real data. One of them is standing on a foundation a journal already flagged. Treat them the same at your own risk.
Written by Ursula Zamora, clinical-topics writer. Reviewing the trials and labels directly. Last reviewed June 2026.
For general readers, not a prescription. Check in with a qualified clinician before you begin.
