What is Semax? What it does, benefits and what the evidence shows

Semax is a synthetic heptapeptide developed in Moscow in the late 1970s and 1980s, built from a fragment of adrenocorticotropic hormone (ACTH) with a stabilizing tripeptide tacked on so it survives long enough to do something. It's included on Russia's official list of vital and essential drugs, and the overwhelming majority of research on it — including every human trial we could verify — comes out of Russian institutions. It's sold in the U.S. grey market as a nasal or injectable "nootropic" for focus, memory and stroke-adjacent recovery, but almost none of that use case has been tested outside Russia, and FDA has never approved it for any purpose. The agency's own listing for Semax is a compounding-nomination screen, not a drug review: its page on bulk substances places Semax under "nominated but withdrawn" (FDA, content current as of April 22, 2026).

What Semax is and how it works

Semax's sequence is Met-Glu-His-Phe-Pro-Gly-Pro — the ACTH(4-7) fragment (the piece of the hormone associated with effects on the brain rather than the adrenal glands) with a Pro-Gly-Pro tail added to slow its breakdown by enzymes. It was developed by a research group led by Nikolay Myasoedov and Ivan Ashmarin at what's now the Institute of Molecular Genetics of the Russian Academy of Sciences, starting from earlier work on ACTH fragments as candidate nootropics (Kolomin et al., Neuroscience and Medicine, 2013). Semax is included in Russia's official "List of Vital and Essential Drugs for Medical Application" (Government of the Russian Federation, Decree No. 2406-r, October 12, 2019), according to a 2025 peer-reviewed paper from the same Moscow research network (Radchenko et al., Acta Naturae, 2025). That's a real, citable regulatory fact — it is not the same thing as FDA or EMA approval, and we found no equivalent Western regulatory status for it anywhere.

In rodents, Semax activates striatal serotonin metabolism and, combined with amphetamine, potentiates dopamine release and locomotor activity (Eremin et al., Neurochemical Research, 2005). Separately, a single administration raised hippocampal BDNF (brain-derived neurotrophic factor) protein and receptor activation in rats, alongside improved performance on a conditioned-avoidance learning task (Dolotov et al., Brain Research, 2006). In rats with induced cerebral ischemia, Semax and a related melanocortin-derived peptide increased blood vessel density and neuroglial proliferation around the damaged area 24 hours later; the authors propose an allosteric effect on receptor systems and note that involvement of melanocortin receptors has not been demonstrated (Stavchansky et al., Current Issues in Molecular Biology, 2024). Those are rat and cell-culture findings. They describe a plausible mechanism; they don't by themselves establish what Semax does in a person.

What people use it for

What the research actually shows

Stroke recovery — the largest human dataset, and still thin by Western standards

The oldest and most-cited human study is a 1997 Russian trial that gave Semax, on top of standard intensive therapy, to 30 patients in the acute period of hemispheric ischemic stroke, comparing them with 80 patients of similar severity and stroke location treated with standard therapy alone. The paper reports Semax-treated patients showed faster regression of general cerebral and focal deficits, particularly motor problems, tracked with clinical rating scales and electrophysiological measures like EEG mapping (Gusev et al., 1997). That's a real signal worth taking seriously, but it isn't a randomized, blinded, placebo-controlled trial — the abstract describes a comparison group treated with standard therapy alone and mentions no placebo, randomization or blinding, and it's a single Russian research group's data from nearly three decades ago. We could not locate a subsequent Western replication of this design. A separate 2024 rat study from the same broader research network cites "the effectiveness of Semax as treatment for patients at different stages of stroke" as established background, but that citation traces back to the same Russian clinical literature, not an independent confirmation (Stavchansky et al., 2024).

The animal work behind that human claim is more recent and more detailed than the human trial itself. A 2020 study induced transient middle cerebral artery occlusion in rats, gave Semax or saline, and sequenced RNA from brain tissue at 4.5 and 24 hours after the injury. At 24 hours, the researchers identified 394 genes with significantly different expression between groups; broadly, Semax suppressed genes tied to inflammatory processes and activated genes tied to neurotransmission, in effect pushing the post-stroke gene-expression pattern back toward what it looked like before the injury, where untreated ischemia pushed it further away (Filippenkov et al., Genes, 2020). That's a detailed molecular story in rats. It supports a plausible mechanism for the older human observations; it isn't itself evidence of what happens in a person's recovery.

Cognition and memory — rat data plus one small human brain-imaging study

Most of the "nootropic" case for Semax is rodent work: hippocampal BDNF and trkB receptor increases alongside better learning on a conditioned-avoidance task in rats (Dolotov et al., 2006), and striatal serotonin and dopamine-system effects, again in rats (Eremin et al., 2005). The closest thing to a controlled human cognitive study we could verify is a 2020 imaging study that gave 52 healthy adults Semax, Selank or placebo and mapped resting-state functional connectivity, finding group differences between the right amygdala and regions of the right temporal cortex (Panikratova et al., 2020). That study shows the peptide changes something measurable in a living human brain; it does not show it improves memory, attention, or any real-world cognitive outcome — no cognitive test scores were reported in the abstract we reviewed.

Alzheimer's-relevant effects — mouse model and test-tube chemistry, not a trial

A 2025 paper tested Semax and a Semax derivative in a transgenic mouse model of Alzheimer's disease and reported "measurable improvements" on standard rodent cognitive tests (open field, novel object recognition, Barnes maze), and separately noted "a preliminary trial of Semax in a limited cohort of AD patients" without giving patient numbers or outcomes we could verify independently (Radchenko et al., 2025). A 2022 chemistry paper, working with artificial membrane models and a human neuroblastoma cell line rather than live animals, found Semax can bind copper and reduce the toxic amyloid-beta aggregation that copper promotes in Alzheimer's-relevant test-tube systems (Sciacca et al., 2022). Both are early-stage findings about a mechanism, not evidence Semax treats or prevents Alzheimer's in people.

The ADHD idea is a hypothesis, not a finding

You'll sometimes see Semax connected to ADHD and Rett syndrome. That traces to a single 2007 theoretical paper that reasons from Semax's known dopamine-potentiating and BDNF-boosting effects in rodents to propose it as worth investigating for these conditions (Tsai, Medical Hypotheses, 2007). A hypothesis paper is exactly what it sounds like: a case for why someone should run a study, not a study. We found no published trial testing Semax in ADHD or Rett syndrome patients.

Side effects and risks

Human safety data for Semax is limited to what's reported inside the same small Russian studies described above, plus the imaging study — none of which was designed primarily to catalog side effects the way a dedicated safety trial would. We found no published randomized safety trial. FDA's page on compounding bulk substances lists Semax under "nominated but withdrawn," saying compounded Semax may pose an immunogenicity risk and that the agency lacks enough information to know whether it would cause harm in humans (FDA, content current as of 04/22/2026); that is a screening note, not a drug review. FDA has never approved Semax, and it carries none of FDA's approved-drug labeling.

That could change on the compounding side. At its July 23–24, 2026 meeting, FDA's Pharmacy Compounding Advisory Committee voted to recommend adding Semax, along with BPC-157, KPV, TB-500, MOTS-C and Epitalon, to the 503A bulks list that licensed compounding pharmacies can draw on. The vote is non-binding: the Secretary of Health and Human Services still has to approve any addition, and pharmacies can't compound it until the final rule is in place (NCPA, July 31, 2026). Even then it would be a prescription, pharmacy-compounded product, not an FDA-approved drug.

On the grey-market side, our data flags a real, practical issue that has nothing to do with the biology: overfill. As of September 29, 2026, Semax vials in our database ran a median of 17.5% over their labeled amount, with the large majority testing more than 10% over and several more than 20% over. Purity was consistently strong (every measured vial at 99%+), so the problem isn't contamination — it's that the amount of peptide in the vial is unreliable relative to what the label says, in one consistent direction. A vendor who can't hit their own label on fill amount is telling you something about their quality control, even when purity looks clean.

What the lab tests show

As of September 30, 2026, our database includes 12 Semax vials from 7 vendors, drawn from 8 lab reports dated January 2025 through October 2025 — a sample that a single vendor's vials can shift noticeably. On purity, 12 of 12 vials measured 99%+, median 99.55%. On dose versus label: median 17.5% over label, 1 of 12 vials within ±5%, with 6 measuring more than 20% over. As of September 29, 2026, no vials had failed our definition (empty, under 95% pure, more than 20% under label, or more than 50% over) — the overfill, however heavy, mostly stayed short of that threshold.

Here's how the ranked vendors for Semax look right now; the table shows current verdicts and flags any vendor whose failed-vial rate is above 5%. With only 12 vials tested for this compound as of September 29, 2026, this table may be short or thin on report counts:

Buying smarter

Check a specific vendor's full Semax record on our Semax results page before trusting a single report — a vendor needs five or more lab reports for this specific compound before we rank them, and the number of reports behind a score matters as much as the score. Treat any COA a vendor posts as a starting point, not proof: how to read a peptide COA and fake peptide COAs cover what to check before trusting a number you didn't see tested yourself.

Given how consistently Semax runs over label in our data, a quantity test on your own vial tells you more than another purity check would — purity has looked clean across the board, dose-versus-label hasn't. How to get a peptide tested walks through ordering independent identity, purity and quantity testing, and whether endotoxin testing is worth adding for a peptide like this one.

If your report comes back a fail, stop using that batch, tell anyone who bought from the same run, bring the report to the vendor, and submit it to VendorVerify so the next buyer has one more data point than you did.

Database figures without a date update automatically from VendorVerify's lab results; dated figures are snapshots. This isn't medical advice. Semax is sold for research use only in the U.S. and has never been approved by FDA; its Russian regulatory status does not extend to the United States.