Finnrick vs. Janoshik vs. VendorVerify: what each one actually measures

People ask "have you checked the Janoshik on that?" and "what's its Finnrick rating?" as though the two questions are interchangeable. They're not, and neither is a question about VendorVerify's score the same thing again. One is a testing lab. One is a vendor-grading platform built on crowdsourced testing. One is us — an aggregator of reports that buyers like you commissioned, from whatever lab they chose. Confusing the three makes it easy to misread what a number is actually telling you.

This isn't a competitive takedown of anyone. We don't test anything ourselves, we don't sell anything, and we'd rather a reader understand what Finnrick and Janoshik are actually good at than pretend we do their jobs better. Where we're citing a specific claim about either of them, it's sourced to their own site as of this writing, dated, and linked.

Janoshik: a testing lab

Janoshik is a laboratory. Someone — a buyer, a vendor, a group-test organizer — sends a physical sample, pays for a specific test, and receives a report back. Janoshik doesn't rank vendors, doesn't publish a leaderboard, and doesn't grade anyone. It answers one question: what did this specific sample contain, measured by HPLC and, depending on the test, mass spectrometry.

Janoshik does publish a searchable log of tests at public.janoshik.com, and every report carries a task number and unique key that can be checked at janoshik.com/verification — see how to verify a Janoshik test for exactly how that works. But that verification tool confirms a report is genuine; it isn't a rating system, and Janoshik doesn't present it as one.

About seven in ten of the reports in our own data come from Janoshik, which is part of why we and plenty of other people in this space treat it as a default reference point. That's a fact about our data, not a ranking of labs — we haven't independently audited Janoshik's methodology or accreditation status, and this article isn't the place to do that.

It's also worth being clear about what a Janoshik result is for, structurally. The person who pays for the test is the client, and the client decides what happens to the result — post it publicly, share it in a group test, or keep it private. Nothing about the lab's own process determines who submits a sample or why. A buyer testing a vial they just purchased and a vendor testing their own product before listing it both go through the identical process, get the identical kind of report back, and the report itself doesn't say which one happened — which is exactly why the source of a report (who commissioned it, and why) matters as much as the report's content. That's why VendorVerify records who paid for each report. Most of ours were commissioned by buyers; a report a vendor paid for is labeled on its page and counts at half weight. The lab work is equally rigorous either way; what differs is who chose the vial, and why.

Finnrick: crowdsourced testing plus vendor grading

Finnrick is a different kind of thing entirely: a platform that collects samples from individuals, sources others directly from vendors, sends them to third-party labs, and turns the results into public vendor ratings. As of this writing, Finnrick's own site states it has published 11,999 test results as of September 29, 2026, with leaderboards broken out by peptide (for example, 378 vendors and 4,367 tests for retatrutide, 262 vendors and 2,744 tests for tirzepatide, as of September 25).

Finnrick's testing options, per its own site:

Finnrick's published methodology describes several named rating standards it applies to that evidence — for example, a "Purity focus" standard requiring at least 99.5% purity and dose within ±35% of label claim, and a "Finnrick Standard" requiring the same purity floor with a tighter ±25% dose band and a minimum labeling score. Paid membership ($9.99/month, per Finnrick's site) unlocks stricter modes — including a 99.8% purity threshold — and original lab PDF downloads. Finnrick's methodology page also describes a recency-weighted scoring approach: it states newer tests count more than older ones, using what it calls an evidence "half-life," and that vendors with only a small number of classifiable tests have their rating pulled toward a peer baseline rather than resting on a handful of results. That's a genuinely different approach from a simple average, and it's worth reading Finnrick's own methodology page if you want the full formula.

The important structural point: Finnrick both runs (or commissions) the testing and produces the public rating from it, for products it or its users sourced. That's a coherent, clearly disclosed model — Finnrick's methodology page states directly that rankings aren't for sale and that results aren't altered when a vendor objects — but it's a different setup from a report a specific buyer commissioned on a specific vial they personally received.

How Finnrick's scoring actually works

Because Finnrick publishes the mechanics in detail, it's worth walking through them rather than just citing the headline numbers. Per Finnrick's methodology page, each test is first classified as a pass, fail, or incomplete under whichever standard is selected; incomplete tests show up in a vendor's history but don't count toward the pass rate. From there, Finnrick calculates what it calls a "natural rating" — closest to the raw pass rate — and a "final rating," which is what actually appears on leaderboards.

The final rating differs from the natural rating in two ways Finnrick discloses. First, recency weighting: newer tests count more than older ones, using an evidence "half-life" Finnrick currently sets at 730.5 days, meaning a two-year-old test counts for roughly half as much as a brand-new one. Second, thin-data adjustment: a vendor with only a handful of classifiable tests has its rating pulled toward a peer baseline for that peptide, rather than letting one lucky or unlucky result stand in for a real track record. Finnrick's own framing is that this "prevents one lucky or unlucky test from looking too certain" — a reasonable problem to guard against in any small-sample rating system.

Finnrick's site also describes a labeling score, used in some of its standards but not others: it runs from 0 to 2, based on whether a batch identifier and batch-specific claims are printed on the container itself (strongest), on a separate physical insert (partial credit), or only on a website or purchase page (weakest, scoring zero). "Purity focus" modes skip the labeling score entirely and grade only compound purity and fill accuracy.

As one snapshot of what this produces: Finnrick's retatrutide leaderboard, dated September 25, 2026 on its own site, showed 378 vendors and 4,367 tests, with the top-ranked vendor at a 93% rating under its purity-focus standard. Its tirzepatide leaderboard for the same date showed 262 vendors and 2,744 tests. These are Finnrick's own published numbers as of that date, not ours, and we'd expect them to keep changing as Finnrick collects more tests.

VendorVerify: an aggregator of independent lab reports

We don't test anything. We don't source samples. We don't choose a lab. What we do is collect reports from independent labs and organize them into a per-vendor picture. Most were commissioned by buyers — individuals who bought a vial with their own money — from whichever lab they chose. When a vendor paid the lab itself, we label the report and count it at half weight.

In our data as of September 30, 2026: 1,529 vials from 67 vendors, tested between March 2023 and September 2026 (by 20 different labs as of September 28, 2026). A vial counts as a fail if it's empty, under 95% pure, more than 20% under its labeled amount, or more than 50% over it. A vendor needs at least 5 reports before we rank it; below that, it's listed but shown as unranked rather than scored. The score itself combines failure rate, purity, dose accuracy, and consistency across reports, and — similar in spirit to Finnrick's approach, though built independently — newer reports count for more than older ones. Each result on a vendor's page shows the lab and, where the report has one, its report key, so it can be checked with the lab itself.

We sell nothing. No vendor pays us, there are no affiliate links, and there's no paid placement — a vendor's score is exactly what their submitted reports say it is.

Here's how the current top-ranked vendors look on VendorVerify:

Why two scores for the same vendor can disagree

If you check a vendor on both Finnrick and VendorVerify, don't expect the numbers to match, and don't treat a difference as evidence either site is wrong. The inputs are genuinely different populations of tests. Finnrick's data comes from its own community-testing pipeline and paid orders, submitted by people using Finnrick specifically. Ours comes mostly from buyers who commissioned a report on their own, through whichever lab they picked, and then chose to submit it to us — which skews toward Janoshik simply because that's who most buyers in this space already use. The fail thresholds differ too: our data counts a vial as failed if it's empty, under 95% pure, more than 20% under label, or more than 50% over; Finnrick's standards use different purity floors and dose bands depending on which of its named standards you're looking at. Two honest, well-disclosed methodologies can reasonably disagree because they're not measuring the exact same set of vials against the exact same bar.

Where each is strongest

Janoshik is strongest when you have a specific vial in hand and want to know what's actually in it. It's a direct answer to a direct question, on your timeline, for your sample. It can't tell you anything about vendor consistency over time, because that was never its job — it evaluates one sample.

Finnrick is strongest when you want an actively growing, standardized dataset with a transparent, published scoring formula and access to free community testing that adds new data continuously. It can also fund your own test through its free tier, which neither of the other two does. Its rating is a summary judgment across the tests it has, using rules it discloses in detail.

VendorVerify is strongest when you want to see what independent testing, mostly paid for by buyers, across many different labs — not one lab's own program — says about a vendor over time, with each result's lab and report key shown so it can be checked. Because we don't run or fund the testing ourselves, we have no way to generate new results on demand for an under-tested vendor; we can only reflect what buyers have already submitted.

A practical scenario

Say you're comparing two tirzepatide vendors. One shows up on Finnrick's leaderboard with a strong rating and hundreds of tests behind it. The other doesn't appear on Finnrick at all but has a handful of buyer-submitted reports on VendorVerify, all recent, all passing. What do you actually know?

You know the first vendor has a large, standardized body of evidence under Finnrick's disclosed methodology, recency-weighted and adjusted for sample size — that's a meaningfully strong signal, precisely because there's a lot of it and the rules for building it are public. You know the second vendor has a smaller, differently sourced set of reports, each one commissioned independently by a different buyer using a lab of their own choosing, which has its own kind of independence — no single testing pipeline, no single relationship, produced every result. Neither absence is proof of anything: a vendor missing from Finnrick's board may simply not have been tested through Finnrick's pipeline yet, and a vendor with only a few VendorVerify reports may simply be smaller or newer to the buyer community that submits to us. The honest answer is that more independently sourced evidence, from more than one place, beats a single number from either site — which is the actual argument for checking more than one of these when it matters to you.

What none of the three can tell you

This applies across all of them, not just us: a lab report, however it was sourced or rated, describes a vial that was tested on a specific day. It doesn't describe the vial you're about to receive, and it says nothing about how to use, dose, or administer any compound — that's outside what any lab report, from any of these three, can tell you. None of this is medical advice, and nothing here is a recommendation to use any vendor named or unnamed. Compounds sold for research use only carry genuine, unresolved uncertainty that no rating system removes.

When to use which

If you have a vial and want it tested, Finnrick's free community tier or a paid order from Janoshik (or another lab) are the actual testing options — VendorVerify doesn't test anything, so we're not on this list for that purpose. If you want to check whether a report you were shown is genuine, Janoshik's own verification tool is the direct path for a Janoshik report specifically. If you want a sense of how a vendor has performed across many buyers and many labs over time, with every result traceable to its lab report, that's what our vendor ratings are for. Reasonable people use more than one of these together — a Janoshik or Finnrick result to check a specific vial, and a broader pattern across sources to judge the vendor.