FFMI: the number that tells muscle from mass
BMI cannot distinguish a bodybuilder from someone carrying excess fat. Fat-free mass index can, and it is the more useful figure for anyone who trains.
CalculatorLean Mass & FFMIOpen →BMI has one well-known failure mode: it treats all mass identically. A heavily muscled athlete and someone carrying substantial excess fat can produce the same BMI, and the index will label both the same way.
Fat-free mass index fixes exactly this, by doing the same arithmetic on lean mass alone.
What FFMI actually is
Take your bodyweight, subtract the fat, and divide the remainder by your height in metres squared. That is FFMI. Structurally it is BMI with the fat stripped out, which means it measures how much lean tissue you carry relative to your frame.
Because it needs lean mass, it needs a body-fat estimate first. This is the main practical limitation — an FFMI derived from a wild guess at body fat is a wild guess at FFMI.
The normalised figure adjusts the result to a 1.8 metre reference height using the correction the original study specified, because taller people tend to produce slightly lower raw values for the same relative muscularity. Normalising makes comparisons across heights more meaningful.
Why it beats BMI for trained people
Consider two men of the same height and weight. One carries 12 percent body fat, the other 30. Their BMI is identical. Their FFMI differs by several points, and it is the FFMI that reflects what is actually going on.
This is the concrete answer to the familiar complaint that BMI misclassifies athletes. The complaint is correct, and FFMI is the fix. It is worth being precise about the scope of that claim, though: BMI remains a reasonable population-level screening tool, and the misclassification problem mostly affects the tails — the notably muscular and the notably sarcopenic. For a sedentary population average, BMI is not as useless as fitness culture sometimes suggests.
Interpreting your value
For men, rough bands look like this: below 18 is below average, 18 to 20 is average, 20 to 22 is above average, 22 to 23 is excellent, and above 23 is well into the territory of serious long-term training. Women’s values run several points lower across the board due to differences in typical lean mass.
Population percentile data exists and is worth knowing about — the most widely used reference set covers 5,635 adults and shows how FFMI distributes across age and sex, which is more informative than the informal bands circulating in fitness communities. 2
Know what that reference set is before leaning on it. The sample was Swiss and Caucasian, selected non-randomly, and body composition was measured by bioelectrical impedance rather than DXA — a method cross-validated against DXA in that study, but still an estimate built on an estimate. Percentiles from it describe that population, not necessarily yours. 2
Three caveats on the measure itself. FFMI does not distinguish muscle from other lean tissue such as bone, organs and water, so it is a proxy rather than a direct muscle measurement. And it says nothing about strength, distribution or health — a high FFMI is not automatically a good thing, nor a low one automatically bad.
The "natural limit" claim
You will frequently see the claim that an FFMI of 25 is the natural ceiling, and that anything above it implies pharmaceutical assistance. This traces to a single 1995 study of 157 male athletes — 83 steroid users and 74 non-users — in which the normalised FFMI of the non-users extended up to a limit of 25.0, while many users exceeded it and some passed 30. A separate group of twenty pre-steroid-era Mr. America winners averaged 25.4. 1
It is worth being careful with this. The sample was small, drawn from a specific population, and the authors themselves described their findings as preliminary and proposed FFMI as an initial screening measure rather than a verdict. 1 They also noted that their non-user sample included only six men above 20 percent body fat, making it hazardous to generalise the ceiling to heavier individuals. Individual variation in frame size, limb length and genetics is substantial, and the measurement error in the body-fat estimate feeding your FFMI is itself several points.
Treat 25 as a rough population observation rather than a diagnostic threshold. Using it to make accusations about specific individuals asks far more of the number than the underlying evidence supports.
The bottom line
FFMI is the better metric if you train, because it measures the thing you are actually trying to change. Calculate it, note it, and re-check it every few months alongside your body-fat estimate.
Just remember it inherits every bit of error from that body-fat figure — and treat the famous ceiling number as a loose observation rather than a verdict.
- Kouri EM, Pope HG Jr, Katz DL, Oliva P. Fat-free mass index in users and nonusers of anabolic-androgenic steroids. Clin J Sport Med, 1995;5(4):223-228.Checked against the original
- Schutz Y, Kyle UUG, Pichard C. Fat-free mass index and fat mass index percentiles in Caucasians aged 18-98 y. Int J Obes Relat Metab Disord, 2002;26(7):953-960.Checked against the original
- Nuttall FQ. Body mass index: obesity, BMI, and health: a critical review. Nutr Today, 2015;50(3):117-128.Checked against the original
How we reviewed this article
Every factual claim here is sourced to peer-reviewed research, a government or institutional guideline, or a named textbook. Where a claim carries a numbered marker, it links straight to that source below, so you can check the original rather than take our word for it.
This article has not been reviewed by a credentialed professional. It was researched and written from the primary sources listed below, but a reviewer catches things citations cannot — omitted contraindications, misread effect sizes, and questions of emphasis. We would rather tell you that than imply a review that did not happen.
Written and edited by a person. All 3 citations have been opened and checked against the original — that the source exists, and that it supports the specific claim it is attached to.
We revisit articles as the evidence changes. This one was last updated on August 20, 2026. Spotted an error? Tell us — corrections are logged publicly.
Estimates and general guidance only. This is not medical advice.
- Corrected which value the interpretation bands are applied to. FFMI categories were being read from raw FFMI rather than the height-normalised figure the published bands were derived from, which mislabelled anyone far from 1.8 m by roughly one band — flattering shorter readers and under-rating taller ones. Categories now use the normalised value.
- Adjusted the top band boundary from 26 to 25, matching the ceiling this article already cited.