Where to buy a qpAdm analysis in 2026, and how much it costs

You cannot buy qpAdm. It is a method inside ADMIXTOOLS 2, a free, open-source R package from the Reich laboratory, and anyone with a laptop and a merged dataset can run it for nothing. What you can buy is one of two things: a done-for-you analysis, where models are tested on your genome and a person approves the one that is published with its numbers, or a do-it-yourself environment, where someone has done the heavy merging and hosts the software so you can run models yourself. Those are different products with different prices, and most of the confusion in this market comes from not saying which one is on offer.

This page lists every service I could find that sells or provides qpAdm in 2026, with the same five facts for each, then sets out what an analysis costs (our four prices and the add-ons, what does not change the price, and when a do-it-yourself subscription is the cheaper route) and ends with a checklist for choosing. I run one of the services, Ancient Ancestry, Ancestrify's qpAdm analysis, so I have kept every competitor entry to what is readable on the competitor's own site and written "not stated on their site" where it is not. Competitor facts were read on each site on 2026-08-30 and may have changed since; our own prices were checked on 2026-09-25.

The two things being sold#

A done-for-you analysis takes your raw file, merges it into a reference panel, tests candidate models against a stated bar, and has a person approve the one that is published. What you receive is a report with a p-value, weights, standard errors and Z-scores, the outgroup set, and a person's approval behind the result. The price is analyst time.

A DIY environment hosts ADMIXTOOLS 2 (or a reimplementation of it) and a reference panel behind a web page. You choose sources and outgroups and press run. What you receive is the raw output; the interpretation, the search and the audit are yours. The price is compute, usually as a subscription or a run allowance.

Neither is better in the abstract. If you already know how to choose a right set and read a nested-model table, a DIY tool is the cheaper route. If you do not, a DIY tool will produce a passing model for you within an hour, and without a reason to expect those sources before the test, the pass tells you little, for the reasons set out in the tutorial and under why a person is in the loop.

Ancestrify#

Every published model clears the same bar, a house convention rather than a validated criterion: p > 0.05, every |Z| > 3 and every SE < 0.10, for every source in every era. With every weight positive, |Z| > 3 places each weight more than three standard errors from 0 and 1. A free Raw DNA File Check tells you before paying whether the file's coverage can support the bar, and the free AdmixTools 2 Lab runs real ADMIXTOOLS 2 over the public panel so you can learn the method first. There is a demo that mirrors a real report. Product page: /qpadm.

SpartaDNA#

spartadna.com describes offering qpAdm analysis. Beyond that description I could not read the accepted inputs, the panel version, the contents of the deliverable or a price on the site on 2026-08-30. If you are considering it, ask for those five facts directly, and in particular whether the complete right set and the standard errors are published with the weights.

Illustrative DNA (AdmixLab)#

Illustrative DNA's AdmixLab is a DIY environment that lists qpAdm and Fst among its tools, with both AADR v62 and v66 available as panels and a stated allowance of up to 100 runs daily. It is a sound place to run your own models if you already know how to choose a right set. What it does not do, by design, is build or audit a model for you.

Genoplot#

Genoplot offers a set of free population-genetics tools and states that qpAdm and f-statistics are among them, with a community forum where users share and discuss models. For someone who wants to learn the method with other people around, it is the friendliest free entry point. It is a DIY route: no one builds or checks a model for you.

qpadm.app#

qpadm.app runs qpAdm client-side, in the browser, on a consumer raw file. Their site lists the accepted vendors and requires an upload code to use it; I could not read where the code comes from, which panel is used, or a price. Running the computation in the browser is a genuine privacy advantage, since the file need not leave your machine. Whether the panel behind it is current enough for your question is something to ask before relying on a result.

Insights on Ancestry#

ioancestry.com is a subscription DIY environment. Their site states two monthly plans, 5.99 USD and 10.99 USD, each carrying a compute balance that runs draw down. It offers qpAdm and Fst; the AADR version behind the panel is not stated on the site. If you plan to run many models over months, the arithmetic of a subscription versus a one-time analysis is worked through below, in what a DIY subscription costs instead.

DNAGENICS#

DNAGENICS belongs in this list mainly to clear up a name. Their site sells Global25 coordinates at 14 EUR, and hosts a calculator called "Mini Qpadam V2" by TheEventTrooper. Despite the name, that calculator is a Global25 coordinate-fitting calculator: it fits a 25-number row to a source panel and returns percentages with a fit distance. It is not qpAdm. It has no f4-statistics, no outgroups, no standard errors and no p-value, and it cannot reject a model. The distinction is the subject of qpAdm vs Global25. Note also that Global25 coordinates issued by anyone other than Davidski's Eurogenes G25 Requests portal are not the official rows that portal produces; see What are Global25 coordinates.

Side by side#

"Tool output" means the numbers are there if you run a model, but nobody has checked the model.

What it costs#

On the web, Ancient Ancestry, the person-approved qpAdm analysis from Ancestrify, costs 29.99, 39.99, 49.99 or 59.99 EUR, paid once, with optional 10 EUR add-ons at checkout and no subscription. A refined re-analysis, offered only if a better model is ever found for your report, is 15 EUR, and only if you choose it. SpartaDNA, the only other done-for-you service above, states no price on its site. A do-it-yourself tool costs nothing or a monthly subscription, and buys a different thing. The rest of this section sets out what each figure buys, what does not move the price, and how to decide which of the two products is the right spend.

The four tiers#

Every tier gets the same report, the same maps and videos, the same model record and the same publish bar. The order page sets the four tiers side by side.

The add-ons#

Four options at checkout cost 10 EUR each:

• Fast compute moves the order to the front of the queue.
• Whole-genome VCF upload is for a sequenced genome (Nebula, Dante Labs, tellmeGen and similar) delivered as a .vcf or .vcf.gz rather than a chip export. The file is converted to the reference panel's markers at upload, lifted from GRCh38 to GRCh37 where needed, and the VCF itself is not stored. Details: Upload a whole-genome VCF.
• Paternal (Y-DNA) haplogroup and maternal (mtDNA) haplogroup are separate products bundled onto the same order, not parts of Ancient Ancestry.

Refined re-analysis, 15 EUR, is never sold at checkout. It is offered only when a better model is found for a published report, as a second version: the original stays exactly as published, you can switch between the two in the report, and you are never charged unless you choose it.

The Ancestry Workbench is not an add-on. It is a prepaid balance from 9.99 EUR that never expires, drawn per completed run, where you run your own models with your own merged genome as the target. A finished report's sample attaches as a kit for free (a kit prepared from a new upload is 0.49 EUR), and the report's Your model view opens the published model there directly. A qpAdm run is about 0.004 EUR, a rotation about 0.004 EUR a model with a 0.04 EUR minimum, and an Fst scan about 0.004 EUR. Failed runs are never charged, and the exact EIGENSTRAT bundle the report was computed from is included, for reproducing everything on your own machine. Page: /workbench; walkthrough: Run your own qpAdm models.

What does not change the price#

Your file's coverage. A 23andMe v5 export and an older AncestryDNA v1 export cost the same to analyse, though they will not support the same standard errors. Coverage bounds what a report can say, not what it costs. The free Raw DNA File Check counts your file's markers before you pay and says whether qpAdm can run on it; how well the chip overlaps the ancient panel is measured during the analysis itself.

Your vendor. 23andMe, AncestryDNA, MyHeritage, FamilyTreeDNA, LivingDNA and similar exports are accepted at the same price.

Your region. The analysis covers every sovereign state, 199 countries and 1,119 regions, at the same price. A genome from a region with sparse ancient sampling is not cheaper to model; it is harder, and that is absorbed.

How many eras. Every report covers both eras (Hunter-Gatherer and Neolithic Farmer; Classical Antiquity) at every tier.

What a DIY subscription costs instead#

The alternative to buying an analysis is renting the environment to do it yourself. The prices are in the directory above: Genoplot's tools are free, Insights on Ancestry states 5.99 or 10.99 USD a month, and Illustrative DNA's AdmixLab does not state its subscription price.

The comparison is not "cheap versus expensive". A subscription is cheaper if you run models every week and already know how to choose a right set, read a nested-model table, and judge whether a passing model means anything. It is more expensive, in both money and hours, if what you want is one defensible answer: a month or two of subscription plus the evenings spent learning the method adds up to roughly the price of a Base analysis, and at the end nobody has checked your model. The free AdmixTools 2 Lab is the right place to find out which kind of user you are before spending on either.

Why a person is in the loop#

The price of a done-for-you analysis is analyst time, so it is fair to ask what that time buys that a DIY run does not. Trying source and outgroup combinations until one passes is the obvious way to sell qpAdm for a few euros, and on its own it proves little: run enough models and one will clear any threshold by chance, and the best-scoring model is frequently not the right one, which is why Harney and colleagues (2021) warn against exactly that practice.

The audit literature's lesson is that a search's error rate depends mostly on its pre-study odds, how plausible its candidates were before any test, not on whether a script or a person ran it. So the work runs in a fixed order. Candidate sources come from a pool curated for your declared origins and era before any model runs. Our analyst tests models from that pool, automated searches included, and re-tests the survivors. Then I review the candidates, choose the model to publish and approve it. You receive one model per era, with its record on screen; the nested models and the rank test are what our analyst checked before publishing, and they are not printed in the report.

Even an approved model describes your genome's ancestry composition relative to the sampled groups; it is not proof that those groups mixed. That order of work, and a record anyone you show it to can check, is what the one-time price pays for.

Refunds#

An analysis is refundable before it runs, and not once the report has been delivered, except where the law requires otherwise; if a fault on our side stops the report being produced in a reasonable time, you can request a full refund. A refund is a request, and a person reads and decides each one. The full policy is at /refund, and the free file check exists so that the main reason for wanting a refund, a file that cannot support the bar, is caught before any money changes hands.

A checklist for choosing#

1. Which of the two things do you want? If you can explain, unprompted, why an outgroup that shares drift with a source breaks the test, a DIY tool will serve you. If not, buy the analysis, and consider a DIY tool afterwards.
2. Is the reference panel named and versioned? AADR v66 is the current release; a product that will not say which panel it uses cannot be compared with anything.
3. Are the p-value, every standard error and every Z-score published beside the weights, and is the right set complete? Without those the weights cannot be evaluated by anyone, and what you hold is not a qpAdm result you can check.
4. Is the rest of the record there for another analyst? The chi-square and degrees of freedom behind the p-value, a confidence interval per source, every panel label and sample count. A model nobody else can check is a claim, not a finding.
5. Does it state a bar before the run? "Good fit" is not a bar. Numbers are.
6. If it sells tiers, does it say what they change? The bar should be the same at every price.
7. Does it tell you what your file's coverage supports before you pay? Check it for free before paying anyone; no product can shrink a standard error the file has fixed.
8. Does it call any population your ancestor? It should not. A source population is a statistical proxy, and a fitting model describes your genome relative to the sampled groups, not a line of descent from the people in them.
9. Does the price make sense for how you will use it? A subscription is cheaper for someone who runs models every week; a one-time analysis is cheaper for someone who wants one answer done properly.

The coordinate and Global25 markets have their own directories: where to buy G25 coordinates and where to buy a G25 analysis. If the answer to the checklist is "the analysis", ours starts at 29.99 EUR at /buy-qpadm-analysis, and the full price list is at /pricing. If you would rather learn first, the free AdmixTools 2 Lab and the demo cost nothing.