Female Testosterone Levels Chart: Normal Ranges and How to Read Your Result

What counts as a normal testosterone level for a woman

A 30-year-old, normally cycling woman has a total testosterone between 15 and 46 ng/dL — 0.5 to 1.6 nmol/L — at the 5th-to-95th percentile, from the widest and most-cited reference-range study of its kind (Braunstein et al., 2011). That range narrows with age: the same concentration that’s unremarkable at 22 can sit above range at 42, because testosterone declines gradually across the reproductive years.

The 5th-to-95th percentile boundary itself is a judgment call, not a law of biology. A narrower 10th-to-90th cutoff would flag more women as abnormal for the same underlying physiology, and a wider 2.5th-to-97.5th cutoff would flag fewer — which is part of why different reference studies, using the same underlying population, can still print slightly different “normal” numbers. The 161-woman cohort behind the numbers below was measured using extraction and chromatography methods rather than a simple direct assay, which is part of why this particular range is so widely cited despite a fairly modest sample size.

What counts as “high” — and why the cutoff moves

A total testosterone reliably above your lab’s stated ceiling — commonly cited somewhere around 45–60 ng/dL, or roughly 1.6–2.1 nmol/L — is the usual definition of high, but that ceiling isn’t fixed the way a blood glucose cutoff is. What’s actually behind an elevated result, and how common each cause is, is covered here; this page stays with the number itself.

The reason the ceiling moves is the assay behind it. Commercial testosterone assays were built and validated for the concentrations found in men, which run far higher than in women — at the low end of the female range, that same assay loses precision fast (Rosner et al., 2007). Two labs running the identical sample through two different assay platforms can report numbers that differ by more than either lab’s stated margin of error.

The free androgen index — the number that actually matters more

The free androgen index (FAI) — total testosterone in nmol/L, divided by SHBG in nmol/L, multiplied by 100 — outperformed total testosterone, calculated free testosterone and calculated bioavailable testosterone as a marker for distinguishing PCOS from normal cycles in a study of 312 women, with a cutoff around 7.1% correctly ruling out 82% of women without the condition (Nadaraja et al., 2018).

Worked example: a total testosterone of 40 ng/dL converts to 1.39 nmol/L. Divided by an SHBG of 20 nmol/L and multiplied by 100, that’s an FAI of 6.95% — just under the cutoff, despite a total testosterone that might already look “high” on its own. Low SHBG is doing most of the work in a result like that, which is exactly why FAI reads the two numbers together instead of testosterone alone.

The same study also tested two other calculated markers, and FAI wasn’t actually the best performer of the three. Calculated bioavailable testosterone (cBT) had the highest area under the curve of any index tested, correctly identifying 83% of women without PCOS at a cutoff of 0.86 nmol/L, with calculated free testosterone (cFT) landing in between at 80% specificity for a cutoff of 0.8 pmol/L. FAI is still the one most labs actually report, for a practical reason rather than an accuracy one: total testosterone and SHBG are already routinely measured, so the arithmetic needs no extra assay or specialised calculation.

Why the same test result reads differently at two labs

Because a direct free-testosterone immunoassay is unreliable at female concentrations, the guideline on this recommends calculating free testosterone from total testosterone and SHBG together, rather than trusting a direct free-T result on its own (Martin et al., 2018). That single choice — calculated versus direct — is one reason your result can look different from a friend’s identical number tested at a different clinic.

The second reason is the reference range itself. A lab builds its “normal” range from its own assay platform and its own local population, then reports results against that specific range — not a universal one. That’s why the Braunstein ranges above and a hospital lab’s printed range can both be correct and still not match exactly: they were built from different assays, different cohorts, and sometimes different percentile cutoffs entirely.

Does the time of day or your cycle day change the result?

Yes — draw timing measurably changes what a testosterone test reports. Androgens run higher earlier in the day, which is why the guideline on testing for elevated androgens specifically recommends a morning blood draw rather than an afternoon one (Martin et al., 2018). Cycle day matters less dramatically than it does for estrogen or progesterone, but it isn’t irrelevant: a sample drawn during an anovulatory stretch — common with PCOS itself — won’t necessarily match one drawn during an ovulatory cycle in the same woman. None of that shows up on a printed reference range, which is built from samples collected under standard, controlled conditions rather than the real variability of an actual cycle.

Who this chart doesn’t work for

These ranges assume a healthy, regularly cycling, premenopausal woman not using hormonal contraception — which doesn’t describe everyone reading this. Pregnancy raises testosterone independent of any underlying condition. Perimenopause and menopause shift the whole range downward well before periods stop altogether. Hormonal contraception raises SHBG substantially, which pushes a calculated free testosterone artificially low against a range like the one above. And these figures don’t transfer directly to adolescents, whose androgen levels are still settling in the years after a first period — a separate, evolving reference problem the chart above doesn’t attempt to cover. If any of that applies to you, the number needs a clinician’s interpretation even more than usual, not less.

Note: PCOS was renamed polyendocrine metabolic ovarian syndrome, or PMOS, in May 2026 by a global consensus of more than 50 organisations (Teede et al., 2026) — same biology, same ranges, only the name on the diagnosis changed. This page uses PCOS, since that’s still what most readers search.

If the picture your labs paint is androgen excess without a lab-catchable cause, that excess doesn’t only show up as a number — it shows up on skin as often as it does on a report, and treating it has its own evidence hierarchy worth knowing before you start anything.

Your next step

Ask specifically for SHBG alongside total testosterone at your next draw, not testosterone alone, so a free androgen index can actually be calculated. Note whether you’re on hormonal contraception when you book, since it raises SHBG and can push a calculated free testosterone artificially low — and bring your cycle history with you, because that timeline tells a clinician more than the number by itself.