What a Biological Age Test Costs, and What You Get for the Money

A biological age test costs between $75 and $500 in the current consumer market. The floor is a blood-chemistry clock built from a standard chemistry panel and complete blood count, advertised around $75 to $100. The middle is a DNA methylation array run on saliva or blood, typically $199 to $349. Above that you find research-grade methylation panels with multiple clock outputs and pace-of-aging estimates, $400 to $600, with a few clinic-bundled products passing $1,000. The price tells you almost nothing about the measurement quality. What matters is the assay, the number of clocks reported, and whether you get the raw data. It also matters whether the test is reliable enough to detect a change in you over time.

What each price tier buys

The $75 to $100 tier is a phenotypic age calculation. These products draw blood, run a metabolic panel and a CBC, and push nine values plus your chronological age through a published regression. The Levine PhenoAge model uses albumin, creatinine, and glucose. It also uses log C-reactive protein, lymphocyte percentage, and mean cell volume. The remaining inputs are red cell distribution width, alkaline phosphatase, and white blood cell count. The lab work costs the vendor perhaps $25 at volume. You are paying for the draw, the order, and the arithmetic.

The $200 to $400 tier is a methylation array. Almost all of it runs on the Illumina Infinium MethylationEPIC BeadChip, which interrogates roughly 850,000 CpG sites in version 1 and about 935,000 in version 2. Vendors advertising “1,000,000+ sites” are usually describing EPIC v2 plus custom content. The reagent and scanning cost is in the low hundreds of dollars. The margin sits in the report.

The top tier adds second-generation and third-generation clocks: GrimAge and its component protein predictors, principal-component versions of Horvath and Hannum, and DunedinPACE. Some add organ-specific or system-specific age estimates derived from the same array. There is no additional wet lab work. One array run supports every clock in the literature, which is why a vendor can add twenty scores at zero marginal cost.

Why the assay matters more than the price

Methylation age estimation from blood is a mature measurement with a decade of validation behind it, and the analytical performance of the array is well characterized ¹. The weak point is the distinction between technical variation and biological variation in the clock output. The chemistry is not the problem.

Run the same DNA sample twice on two arrays and first-generation clocks can disagree by a year or more. Sehgal and colleagues quantified this directly by comparing technical replicates against biological signal across a panel of clocks, and found that reliability varies enormously between clocks built on the same data ². Principal-component reconstructions of the classic clocks are substantially more reliable than the originals because they average across correlated probes instead of trusting a few hundred individual CpGs. If you are buying a test to track a change, buy one that reports PC-based clocks or DunedinPACE and ignore the raw Horvath number.

This has a practical consequence for how you read a single result. A report saying your biological age is 41.3 against a chronological age of 46 is reporting a difference that may be within the measurement error of the assay. The scores worth attention are the ones with published age norms and known test-retest behavior. Bourassa and colleagues published population norms for DunedinPACE, which lets you place your value on a distribution rather than against a single number ³.

Pace of aging versus age acceleration

Two different quantities get sold under the same label. Age acceleration is the residual of predicted age on chronological age: a snapshot of how old your tissue looks. Pace of aging is a rate, calibrated against longitudinal change in organ-system function, and expressed as years of physiological decline per calendar year. DunedinPACE is the best-known example and uses fewer than 200 CpGs.

Rate measures are more useful for anyone planning repeat testing, because a rate is the thing an intervention would plausibly move. Newer pace estimators trained in independent cohorts, such as FraminghamPACE, extend this approach to different populations and follow-up structures ⁴. Age acceleration measures are more useful as a one-time cross-sectional signal. They have detected group differences in clinically distinct populations, including elevated acceleration in young adults with congenital heart disease ⁵.

Be skeptical about intervention response. A pilot substudy of the REPRIEVE trial examined whether pitavastatin shifted epigenetic aging biomarkers in people with HIV, and the effects on clock measures were limited ⁶. That is the general pattern in the literature so far. Population-level associations with lifestyle and diet exposures are detectable in large samples, as in the NHANES analysis of oxidative balance score and epigenetic clocks ⁷. Detecting the same effect in one person over six months is a different statistical problem.

Insist on raw data before you pay

The single question that separates a useful purchase from a subscription to a dashboard: do you get the IDAT files? Every Illumina array produces a pair of binary files per sample, one green channel and one red, named by sentrix barcode and position. They are a few megabytes each. If a vendor will not release them, you have bought a number you cannot audit.

With IDATs in hand, the pipeline is short. In R, sesame::openSesame() handles the usual path. That path is background correction with noob, dye bias correction, and detection p-value masking through pOOBAH. The alternative is minfi with preprocessFunnorm(), which is preferable when you have many samples and want functional normalization across them. Either gives you a beta value matrix, where each entry is the fraction of methylated signal at a CpG site, bounded between 0 and 1.

From there, methylclock computes Horvath, Hannum, PhenoAge, and several others in one call. DunedinPACE has its own package and requires the specific normalization the authors used, so do not substitute your own. Compute cost for all of this is a few minutes on a laptop.

Three failure modes account for most bad results. First, EPIC v2 renamed probes with suffixes like cg00000029_BC21, and legacy clock coefficient tables key on the bare identifier. Strip the suffix and collapse duplicates before scoring, or your clock will silently drop half its probes. Second, saliva and blood have different cell compositions, and clock outputs shift with cell type fraction. Estimate leukocyte proportions with estimateCellCounts2() against the FlowSorted.Blood.EPIC reference and report them alongside the score. Third, a high fraction of failed probes from degraded DNA will pass through many pipelines without warning, so check the detection p-value failure rate before you trust anything.

How we would spend the money

If you want one number once, the $75 phenotypic age from a blood panel is fair value, and you can recompute it yourself from any CBC and chemistry panel you already have. If you want to track change, buy a methylation array that returns IDATs and reports DunedinPACE and PC-based clocks, and expect to spend $300 to $400 per timepoint.

Test no more often than every six months. Nothing in the reliability data supports resolving change on a shorter interval ². Draw at the same time of day, from the same sample type, and ideally through the same lab. Switching platforms mid-series introduces a batch effect larger than most real signal.

Finally, treat the clock as one axis among several. Microbiome composition carries independent information about systemic aging, and intestinal microbial signatures have been linked to accelerated aging measures in ways that methylation does not capture ⁸. Combining host and microbiome features improves aging characterization over either alone ⁹. A biological age number is a summary statistic. If a result concerns you, the next step is a clinician with your actual labs. Another test kit is not the answer.

Questions people also ask

Why do two biological age tests give me different answers? They are running different models, often on different tissue, and reporting different quantities. A blood-chemistry clock and a methylation clock are not measuring the same thing, and even two methylation clocks trained on the same array can disagree because of differing reliability ².

Is a $99 test worse than a $400 test? Not necessarily on assay quality, since most sit on the same Illumina platform. The $400 tests generally add pace-of-aging estimates and PC-based clocks, which are the outputs most worth having. Check what clocks are reported before comparing prices.

Does insurance cover it? No. These are consumer wellness products and are not billed to insurance. The underlying blood panel in a phenotypic age test may be covered if a clinician orders it for a clinical reason.

Can I compute biological age from labs I already have? Yes, for phenotypic age. If you have the values below from the same draw, the published PhenoAge coefficients give you the score in a few lines of code.

• albumin
• creatinine
• glucose
• CRP
• lymphocyte percentage
• MCV
• RDW
• alkaline phosphatase
• WBC

What does a DunedinPACE value near 1.0 mean? It means roughly one year of physiological decline per calendar year, which is the cohort average by construction. Published age norms let you see where your value falls relative to others at your chronological age ³.

Footnotes

1. Cordulle Vanthillo, Bram Bekaert. Epigenetic Age Estimation from Blood: Advances, Applications, and Forensic Implications. Epigenetics and Human Health, 2026. https://doi.org/10.1007/978-3-032-23085-0_5 ↩

2. Raghav Sehgal, Daniel S. Borrus, John Gonzalez, et al. Biological Versus Technical Reliability of Epigenetic Clocks and Implications for Disease Prognosis and Intervention Response. Aging Cell, 2026. https://doi.org/10.1111/acel.70635 ↩ ↩² ↩³

3. Kyle J. Bourassa, Calen P. Ryan, Karen Sugden, et al. Age norms for DunedinPACE: An epigenetic pace of aging biomarker. 2026. https://doi.org/10.64898/2026.08.13.26360306 ↩ ↩²

4. William T Marella, Calen P Ryan, David Corcoran, et al. An epigenetic speedometer to measure Pace of Aging: FraminghamPACE. 2026. https://doi.org/10.64898/2026.07.07.26357388 ↩

5. Parag N. Jain, Beryl C. Zhuang, Joanne Whitehead, et al. Epigenetic age acceleration in young adults with congenital heart disease. Clinical Epigenetics, 2026. https://doi.org/10.1186/s13148-026-02049-5 ↩

6. Michael J Corley, Maya Watanabe, Alina P S Pang, et al. Effect of Pitavastatin on Epigenetic Aging Biomarkers in People With HIV: Pilot Substudy of the REPRIEVE Trial. Clinical Infectious Diseases, 2025. https://doi.org/10.1093/cid/ciaf247 ↩

7. Mingming Lv, Zhihui Jiang, Changjiang Deng, et al. Association between the oxidative balance score and biological aging measured by epigenetic clock: insight from the NHANES 1999-2002. Clinical Epigenetics, 2026. https://doi.org/10.1186/s13148-026-02141-w ↩

8. Shalini Singh, Leila B. Giron, Maliha W. Shaikh, et al. Distinct intestinal microbial signatures linked to accelerated systemic and intestinal biological aging. Microbiome, 2024. https://doi.org/10.1186/s40168-024-01758-4 ↩

9. Karina Ratiner, Suhaib K. Abdeen, Kim Goldenberg, et al. Utilization of Host and Microbiome Features in Determination of Biological Aging. Microorganisms, 2022. https://doi.org/10.3390/microorganisms10030668 ↩