Semax is a compound with a real clinical literature and no dose you can carry to a vial. The studies exist, some of them in patients, and they describe what was given — but they describe it as a percentage of a nasal solution and a number of procedures, not as a quantity of peptide. Meanwhile the world's largest trial registry has no record of it at all. This page reports both, with what we opened and when. It does not recommend anything. Our guide to reading a dosing claim sets out what separates a label figure from a convention. How brands appear on this site is set out on our disclosure page.
What it is
Semax is a synthetic heptapeptide with the sequence Met-Glu-His-Phe-Pro-Gly-Pro — the ACTH(4-7) fragment with a Pro-Gly-Pro tail added to slow enzymatic breakdown. It carries no adrenocorticotropic activity of its own and is described in the literature as a neuroprotective and nootropic agent. There is no FDA-approved product, so there is no labelled dose.
Two numbers that describe the evidence
PubMed held 231 Semax records on 2026-09-08, of which exactly one is tagged as a randomized controlled trial. A ClinicalTrials.gov v2 API query for interventional studies naming Semax on the same day returned no studies at all.
That combination is unusual. Compounds with no registry presence normally have no clinical literature either; Semax has decades of one. The research was done inside a system that did not register trials the way the international registries expect, so a reader searching the registry finds nothing and a reader searching PubMed finds hundreds of papers, most of them in Russian and most of them in animals.
FDA reached the same place from the other direction. In its briefing document for the July 23-24, 2026 Pharmacy Compounding Advisory Committee meeting, the agency wrote: "We were not able to find pharmacokinetic studies in humans following exposure to semax (free base) or semax acetate via any ROA," and proposed not adding semax to the 503A Bulks List. The committee voted 8-5 to recommend adding it anyway. Our page on cognitive peptides covers that meeting and the rest of the nootropic-peptide record in full.
A missing pharmacokinetic literature is precisely why the dose figures below cannot be converted. Pharmacokinetics is the study that tells you what a given quantity by a given route produces in the body; without one, a percentage on a bottle and a milligram on a vial have no established relationship to each other.
What the human studies actually gave
These are the dose-stating human records we opened on 2026-09-08.
| Study | What it states | Participants | Link |
|---|---|---|---|
| Vestn Oftalmol 2026;142(3):29-37 | Group 1: endonasal electrophoresis with 1% Semax solution, 10 minutes once daily, course of 8 procedures. Group 2: nasal instillations of 0.1% Semax from day 5 after surgery | 42 (23 and 19) after retinal-detachment surgery | PMID 42366656 |
| Bull Exp Biol Med 2018;165(5):653-656 | Measurements taken directly before and 5 and 20 minutes after intranasal 1% Semax | 14 subjects | PMID 30225715 |
| Dokl Biol Sci 2020;490(1):9-11 | Resting-state fMRI before, 5 and 20 minutes after injection of Semax, Selank or placebo | 52 healthy participants | PMID 32342318 |
| Eksp Klin Farmakol 2015;78(12):30-3 | 0.1% semax solution intranasally, "600 mg/day for 10 days" added to conventional therapy | 60 of 118 patients with psoriasis and metabolic syndrome | PMID 27051926 |
The first three are consistent with each other and with the way the compound is described clinically: a nasal solution at 0.1% or 1%, given as drops or by electrophoresis, in short courses.
The fourth is not usable. At 0.1%, 600 mg of peptide is 600 mL of solution, which is not something anyone instils into a nose over a day. The abstract states it, so we report it, and we say plainly that it is an unresolved unit error rather than a dose figure. That is the same discipline our guide to reading a dosing claim applies to every number on this site: a figure printed in a source is not automatically a fact about a dose.
Why the percentages will not convert
A 1% solution is 10 mg of peptide per mL. That looks like enough to do the arithmetic, and it is not, for two reasons the sources themselves supply. The abstracts do not state the instilled volume, so the mass delivered per session is unknown. And the route is not the route a research vial is sold for — the protocol page that publishes injectable Semax figures states that subcutaneous Semax is "much less common in the published literature than the intranasal route", that human subcutaneous research data is "sparse", that its injectable ranges are "community research planning, not formal trials", and that planning should rely on vial-concentration arithmetic rather than transferring intranasal numbers across routes (Peptide Dosing Protocols, Semax guide, last reviewed August 2026, opened 2026-09-08).
So the honest position is that the published human figures and the figures used with a research vial are not the same kind of number, and no reliable bridge between them exists in any source we opened. Our reconstitution page has the arithmetic for turning a vial and a diluent volume into a concentration, which is as far as arithmetic can take this one.
What is not known
What mass of Semax the published nasal courses delivered; whether any injectable figure has a basis beyond convention; and what either does over a longer period than the ten-day and eight-session courses the literature describes. Nothing on this page is a recommendation.
Sources and dates
Opened 2026-09-08: PubMed record counts and abstracts through the NCBI E-utilities (PMIDs 42366656, 30225715, 32342318, 27051926); the ClinicalTrials.gov v2 API interventional-study query for Semax, which returned no studies; the Peptide Dosing Protocols Semax guide, last reviewed August 2026. The FDA briefing-document quotation and the committee vote were verified on this site's cognitive peptides page, sourced there to the FDA briefing document opened 2026-09-05 and the vote tally reported by RAPS on 2026-07-24. Corrections go to the contact page.