BMI
Body Mass Index and ideal weight ranges
BMI is weight divided by height squared. Adolphe Quetelet devised the ratio in the 1830s as a way of describing populations, and it was adopted as a clinical screening index in the 1980s. Nothing in it measures fat, and it was never designed to describe an individual.
This calculator returns three things: the index, the WHO category it falls in, and — more useful than either — the weight range at your height that corresponds to a BMI between 18.5 and 24.9. At 5 ft 9 in that band runs from about 125 lb to about 169 lb. If you have been treating BMI as a single target, that 43-pound spread is the correction.
Enter a sex alongside height and this also returns an Ideal Weight panel — four classic clinical dosing formulas (Devine, Robinson, Miller, Hamwi), each producing a single number rather than a range. They exist for a different purpose than BMI: estimating a reference body weight for medication dosing, not describing what someone should weigh.
This is a screening number, not a diagnosis. Interpreting it belongs with a clinician who can see the rest of the picture.
The three outputs
- BMI, to one decimal place.
- Category, using the WHO cut points: under 18.5, 18.5 to 24.9, 25.0 to 29.9, and 30.0 or above.
- Healthy Weight Range — the weights at your entered height that produce a BMI of 18.5 and 24.9, converted back into whichever units you selected.
The range line is the one worth writing down. A ratio is easy to chase and hard to interpret; a weight band tells you how much room a single category actually contains, which is usually far more than people expect.
WHO cut points, and where they are contested
| BMI | Category | Population-level association |
|---|---|---|
| Below 16.0 | Severely underweight | Elevated risk of nutritional deficiency |
| 16.0 - 18.4 | Underweight | Elevated risk of nutritional deficiency |
| 18.5 - 24.9 | Normal weight | Lowest average risk band in most cohort studies |
| 25.0 - 29.9 | Overweight | Moderately elevated risk |
| 30.0 - 34.9 | Obesity, class I | Elevated risk |
| 35.0 - 39.9 | Obesity, class II | High risk |
| 40.0 and above | Obesity, class III | Very high risk |
Three qualifications attach to those lines. WHO's 2004 expert consultation noted that many Asian populations show elevated cardiometabolic risk below a BMI of 25 and proposed additional public-health action points around 23 and 27.5. The American Medical Association adopted policy in 2023 stating that BMI should not be used alone to diagnose obesity, citing reference data drawn largely from non-Hispanic white populations and a loss of predictive value at the individual level. And waist circumference — above 40 in (102 cm) for men, 35 in (88 cm) for women — carries information about visceral fat that a height-weight ratio cannot see.
For a fuller treatment, read Why Your BMI Number Might Be Misleading You.
The formula, and why the exponent is arguable
- Metric:
BMI = kg / m^2(centimetres are divided by 100 first) - Imperial:
BMI = lb / in^2 x 703
The 703 is purely a unit conversion; both routes give the same number for the same body.
The squared term is the weak point. Adult body mass scales with height closer to the power 2.5 than 2.0, so the index reads systematically high for tall people and low for short people at identical body composition. A 6 ft 4 in adult and a 5 ft 2 in adult built to the same proportions do not receive the same BMI, and the taller one is penalised for it. That is a property of the formula, not of the people it is applied to, and no amount of careful measurement fixes it.
Four people this number describes badly
The muscular. BMI counts kilograms without asking what they are made of, so a lean athlete and a sedentary person of the same height and weight get the same result. Body fat percentage or a waist measurement is the right instrument: Body Fat Percentage.
Older adults. Muscle is lost and fat gained with age at an unchanged scale weight, so a stable BMI across two decades can conceal a substantial change in composition.
Anyone whose fat is mostly visceral. Two people at BMI 27, same height, same weight, can carry very different metabolic risk depending entirely on where the tissue sits.
Children and teenagers. The same formula is used but read against age- and sex-specific percentile curves rather than these fixed cut points. This calculator applies adult cut points and is intended for adults aged 20 and over.
BMI flags people worth measuring further. It does not diagnose anyone, and a result in the overweight band is a prompt for a second measurement, not a conclusion.
Two worked calculations, and how much rounding matters
- Imperial default: 154 lb at 69 in.
154 / 69^2 = 154 / 4,761 = 0.032346;x 703 = 22.7. Category: normal weight. Healthy range at 69 in: 125.3 lb to 168.6 lb. - Metric default: 70 kg at 175 cm.
1.75^2 = 3.0625;70 / 3.0625 = 22.9.
Those two people are not quite the same person — 154 lb is 69.9 kg and 69 in is 175.3 cm — which is why the answers differ by 0.2. More usefully: at this weight, one inch of height is worth about 0.7 BMI points. Half an inch of measurement error, which is easy to accumulate measuring yourself against a doorframe, moves the result by roughly 0.3. Treat the first decimal as noise and the category boundary as fuzzy.
Where BMI is the right tool
It is cheap, needs no equipment beyond a scale and a tape, and is highly reproducible, which is exactly what a population-level screen requires. As a first pass on yourself it is reasonable, and as an input to a conversation with a clinician alongside waist circumference, blood pressure, lipid panel and family history it earns its place.
It is a poor single metric for an athlete, a pregnant person, or anyone trying to work out whether a change in scale weight came from fat or from muscle. For that pairing, Body Fat Percentage tracks composition and Calorie (TDEE) handles the energy side. The Ideal Weight panel on this page shows how much narrower the classic clinical formulas are than the healthy BMI band at the same height — a single number instead of a 43-pound range.
Frequently asked questions
Is BMI a measure of body fat?
No. It is a height-weight ratio that correlates with body fat across large groups and loses that reliability for individuals. For a direct estimate, use the body fat percentage calculator.
Does BMI mean something different for men and women?
The formula and the WHO thresholds are identical. Interpretation differs, because women carry more essential fat than men at the same index, which makes BMI an even rougher proxy for body fat in women.
Why does my BMI change when I round my height?
At 154 lb, one inch of height moves BMI by about 0.7 points. That is large relative to the width of a category boundary, so measure carefully and do not treat the first decimal as meaningful.
Are the cut points the same everywhere in the world?
Not in practice. WHO's 2004 expert consultation proposed additional public-health action points at 23 and 27.5 for Asian populations because cardiometabolic risk rises at lower BMI values in those groups.
Is BMI valid for children and teenagers?
Not with these thresholds. Paediatric BMI is interpreted against age- and sex-specific percentile curves, because body composition changes rapidly during growth. This tool applies adult cut points.
My BMI says overweight but I lift weights. What should I do with that?
Measure something else. Muscle is denser than fat, so trained people are routinely misclassified. Waist circumference and body fat percentage will both tell you more than the index does.
What are the Devine, Robinson, Miller, and Hamwi formulas in the Ideal Weight panel?
Four clinical formulas from 1964–1984, originally created to estimate a reference body weight for medication dosing, not a personal weight target. Each computes a base weight for 5 feet of height plus an increment per inch over that, differing by sex. They typically produce a single number well inside the wider healthy BMI weight range.
Sources and further reading
Figures and definitions on this page are drawn from the following primary sources. If you find something out of date, tell us and we will correct it.