Overweight starts at 25.0 and obesity at 30.0, the thresholds the WHO and the CDC publish for adults of 20 and over.
Enter a height and a weight and the marker will land on the band.
Measure height without shoes and weight first thing in the morning, because a full day of food and fluid moves the answer by a kilogram or two. BMI is a screening measure, not a diagnosis, and the band it lands in is the start of a conversation rather than the end of one.
Two sets of thresholds for one number
The arithmetic never changes. The line drawn through it does. A WHO expert consultation reported in The Lancet in January 2004 that the proportion of Asian people at high risk of type 2 diabetes and cardiovascular disease is substantial well below a BMI of 25, and set out further public health action points at 23.0, 27.5, 32.5 and 37.5. The NHS applies the first two of those to South Asian, Chinese, other Asian, Middle Eastern, Black African and African-Caribbean family backgrounds. Switch the scale above and the whole band redraws.
| Band | International | Asian and South Asian | What moves |
|---|---|---|---|
| Underweight | Below 18.5 | Below 18.5 | Identical on both scales |
| Healthy weight | 18.5 to 24.9 | 18.5 to 22.9 | The window narrows by two whole points |
| Overweight, or increased risk | 25.0 to 29.9 | 23.0 to 27.4 | Action starts two points earlier |
| Obesity class 1, or high risk | 30.0 to 34.9 | 27.5 to 32.4 | The obesity line falls from 30 to 27.5 |
| Obesity class 2, or higher risk | 35.0 to 39.9 | 32.5 to 37.4 | Same 5 point step, shifted down |
| Obesity class 3, or highest risk | 40.0 and above | 37.5 and above | Severe obesity, the top of both scales |
| Children and teenagers, 2 to 19 | Percentile, not BMI | Percentile, not BMI | Overweight at the 85th, obesity at the 95th |
International bands read on 23 September 2026 from the CDC adult BMI calculator, which lists the three obesity classes separately. The lower thresholds read the same day from the NHS obesity page, and the action points from The Lancet volume 363, issue 9403, pages 157 to 163, 10 January 2004. Child percentiles from the CDC child and teen BMI categories page, which measures against the CDC 2000 growth charts and calls 120 percent of the 95th percentile, or 35 kg/m2, severe obesity.
Which way the error runs
Everybody says BMI ignores muscle. Almost nobody says which direction that pushes your number.
What body mass index knows about you, and what it does not
How to read the band rather than the number, why two different sets of thresholds exist, and which direction the error runs for the people BMI fits worst.
Body mass index is one division. Take a weight in kilograms, divide it by a height in metres multiplied by itself, and the answer is a number that for almost everybody sits between about 12 and about 60. Everything else about BMI is interpretation, and interpretation is where this BMI calculator spends its effort. Instead of printing the number and a single word beside it, the tool draws the whole scale to size and puts a marker where you land, because a BMI of 19.2 and a BMI of 24.8 are both called healthy weight and they are not remotely the same position.
The formula came from Adolphe Quetelet, a Belgian astronomer and statistician who worked it out between 1830 and 1850 as part of a project about the average man. He was describing populations, not patients, and he never meant it for medical assessment. The name arrived a century later. Ancel Keys coined the term body mass index in the Journal of Chronic Diseases in July 1972, and his case for it was deliberately modest: it was, he wrote, "if not fully satisfactory, at least as good as any other relative weight index". Nothing since has made it better than that, and this page treats it accordingly.
It is a screening measure, not a diagnosis. The CDC states plainly that BMI is one potential health indicator and is not intended to diagnose disease or illness. It tells you which conversation is worth having. It does not finish that conversation for you.
Read the band, not the word
The band above is drawn on a BMI axis running from 14 to 42, and every category is as wide as the span it actually covers. Underweight, which runs all the way down, is clipped at the left edge. Healthy weight is six and a half points wide on the international scale. Overweight is five. Each obesity class is five more. Seeing them at true width does something a table cannot: it shows that the jump from healthy weight to overweight is a smaller step than the width of either band, so a reader sitting at 24.8 is one large meal and a dehydrated morning away from a different label. That is not a reason to panic, it is a reason to stop treating the label as a fact about your body.
Under the band the tool prints three proportional readings instead of a verdict. The first is the healthy weight window for your exact height, in kilograms or pounds depending on which tab you used. The second shows where your weight sits inside that window, which is the thing a single BMI figure hides completely. The third is the distance to whichever edge is nearest. Someone at the very bottom of the healthy window and someone at the very top get the same word and very different bars.
Two sets of thresholds for one number
The international scale is the familiar one. Below 18.5 is underweight, 18.5 to 24.9 is healthy weight, 25.0 to 29.9 is overweight, and obesity is split into class 1 at 30.0 to 34.9, class 2 at 35.0 to 39.9 and class 3, also called severe obesity, at 40.0 and above. The WHO publishes the same cut-offs at 25 and 30, and adds sub-bands below 18.5 for severe thinness under 16, moderate thinness from 16 to 17 and mild thinness from 17 to 18.5.
The second scale exists because those lines were drawn against European bodies. A WHO expert consultation published in The Lancet on 10 January 2004, volume 363, issue 9403, pages 157 to 163, reviewed the evidence that Asian populations carry a different relationship between BMI, body fat percentage and health risk, and concluded that the proportion of Asian people at high risk of type 2 diabetes and cardiovascular disease is substantial at BMIs below 25. It stopped short of replacing the international cut-offs and instead identified further public health action points at 23.0, 27.5, 32.5 and 37.5. The NHS applies the first two of those, telling people from South Asian, Chinese, other Asian, Middle Eastern, Black African and African-Caribbean family backgrounds that 23 to 27.4 is overweight and 27.5 or higher is obese, because those groups face health problems at a slightly lower BMI.
Switch the scale selector above and watch what happens. The number does not move. The healthy window closes two points earlier, the healthy weight range in kilograms shrinks for the same height, and a reading in the low twenties can cross from healthy weight to increased risk without a gram changing. That is the most important thing on this page, and almost nothing else ranking for this keyword shows it.
Which direction your number is wrong
Everybody repeats that BMI does not account for muscle. Far fewer sources say which way that pushes the answer, and the direction is the part that matters. For a muscular athlete BMI reads too high, because lean mass is denser than the fat it replaced. As a test for obesity the index has poor sensitivity, catching around 36 percent of men and 49 percent of women who actually have it. For an older adult it reads wrong in both directions at once: height lost with age raises BMI without a gram gained, while muscle lost with age raises true body fat at an unchanged BMI. Cleveland Clinic simply excludes anyone over 65 from the standard chart, along with pregnant people, children, athletes and anyone with muscle atrophy. After the menopause the error runs the other way, with body composition shifting toward fat at an unchanged weight, so a moderate BMI can sit over a body fat percentage that is not moderate.
Children are a different calculation
For ages 2 through 19 the adult bands are not conservative, they are meaningless. A child's BMI is compared with other children of the same exact age and sex on a growth chart and reported as a percentile: below the 5th is underweight, the 5th to below the 85th is healthy weight, the 85th to below the 95th is overweight, and the 95th or above is obesity. Severe obesity is 120 percent of the 95th percentile or a BMI of 35 or more. The CDC measures against its 2000 growth charts, extended in 2022 above the 97th percentile. A BMI of 19 can be a low percentile at one age and a high one two years later, which is why this page refuses to place an entered age under 20 on the band at all.
Pregnancy uses the BMI you had before
Every piece of gestational weight gain guidance is keyed to pre-pregnancy BMI, never to a BMI taken during pregnancy, and no BMI calculator in the ranking set says so. For a single baby the CDC recommends gaining 28 to 40 lb from a pre-pregnancy BMI below 18.5, 25 to 35 lb from 18.5 to 24.9, 15 to 25 lb from 25.0 to 29.9, and 11 to 20 lb from 30.0 to 39.9. If you are pregnant now, the useful number is the one from before, and the band above will not describe you.
The arithmetic, including the part nobody prints
In metric the formula is weight in kilograms divided by height in metres squared. In pounds and inches every source prints BMI = 703 × lb ÷ in², and 703 is a rounding. A pound is exactly 0.45359237 kg and an inch is exactly 0.0254 m, so the true constant is 0.45359237 divided by 0.0254 squared, which is 703.0696. That gap is about one part in ten thousand, invisible at one decimal place, and it is still the reason this tool converts pounds and inches to kilograms and metres first rather than multiplying by 703. The metric tab and the imperial tab return an identical number for an identical body, and the working box under the band shows both so you can see the size of the difference.
The exponent is arguable too. Squaring height was a convenience, not a finding, and work on the relationship between the index and measured fat percentage suggests an exponent nearer 2.5 fits better, which is the basis of Nick Trefethen's proposed 1.3 × kg ÷ m^2.5. Practically this means BMI runs a little high for tall people and a little low for short people. BMI Prime, simply BMI divided by 25, is another way of reading the same figure, where anything at or above 1.0 is over the international line.
- 1Measure a waist as well
BMI knows nothing about where fat sits, and where it sits is most of the risk. The NHLBI threshold is more than 35 inches, which is 88 cm, for women and more than 40 inches, which is 102 cm, for men. Two people with the same BMI and forty centimetres of difference between their waists are not in the same position.
- 2Try both threshold scales
If your family background is South Asian, Chinese, other Asian, Middle Eastern, Black African or African-Caribbean, the lower scale is the one the NHS would apply to you. Run it either way and note whether the label changes, because that difference is a real change in when action is advised, not a rounding.
- 3Watch the direction, not the point
One reading is a snapshot of a morning. The same measurement taken under the same conditions every few months is a trend, and a trend is the only thing in this calculation that carries information about you rather than about a population.
- 4Take it to somebody who can measure more
Blood pressure, glucose, cholesterol, a waist measurement and a look at what you actually eat and do together answer questions BMI cannot. The number above is the opening line of that appointment, not the conclusion.
If what you want next is an energy target rather than a category, the calorie calculator handles resting metabolic rate, activity level and daily calorie goals, which this page deliberately does not.
Frequently Asked Questions
Why does overweight start at 23 for some people and not until 25 for others?
Because the 25 line was drawn against European bodies and does not transfer. A WHO expert consultation reported in The Lancet on 10 January 2004 that Asian populations carry more body fat, and more of it around the abdomen, at any given BMI, so type 2 diabetes and cardiovascular risk climb well before 25. It declined to replace the international cut-offs outright and instead set further public health action points at 23.0, 27.5, 32.5 and 37.5. The NHS applies the first two, treating 23 to 27.4 as overweight and 27.5 or higher as obese for people from South Asian, Chinese, other Asian, Middle Eastern, Black African and African-Caribbean family backgrounds. Use the scale selector above to see both readings for the same number.
My BMI says obese but I lift weights. Which number is wrong?
The BMI is. The index divides total weight by height squared and has no way to tell muscle from fat from bone, so lean mass added through training raises the number exactly as fat would. Measured against actual body fat, BMI is a poor test for obesity, identifying roughly 36 percent of men and 49 percent of women who have it. If you carry visible muscle, a low waist measurement and a BMI in the overweight or class 1 obesity range together tell you the index is reading your training rather than your risk. What it cannot do is prove the reverse, so a high BMI plus a waist over 40 inches for a man or 35 inches for a woman is still worth acting on.
Should I use my pre-pregnancy BMI or the one I have now?
The one from before. Every gestational weight gain recommendation is keyed to pre-pregnancy BMI, and a BMI calculated during pregnancy describes a body that is meant to be changing, so it carries no useful meaning. The CDC recommends a total gain for a single baby of 28 to 40 lb from a pre-pregnancy BMI below 18.5, 25 to 35 lb from 18.5 to 24.9, 15 to 25 lb from 25.0 to 29.9, and 11 to 20 lb from 30.0 to 39.9. Work out your BMI from your weight before conception, read the matching range, and treat the band on this page as not applying until afterwards.
Why will this calculator not give my fourteen year old a category?
Because the adult bands would be wrong rather than merely approximate. Between ages 2 and 19 a BMI only means something compared with other children of the same exact age and sex, reported as a percentile on a growth chart. Below the 5th percentile is underweight, the 5th to below the 85th is healthy weight, the 85th to below the 95th is overweight, and the 95th or above is obesity, with severe obesity at 120 percent of the 95th percentile or a BMI of 35. The CDC measures against its 2000 growth charts. The same BMI of 19 can sit at a low percentile at one age and a high one two years later, so printing an adult label over it would be actively misleading. Enter an age of 20 or more, or leave the age field blank, and the band appears.
Does BMI still mean the same thing after 65?
No, and it goes wrong in two directions at once. Height lost through ageing raises BMI without any weight being gained, which makes the number look worse than the body. Muscle lost through ageing raises real body fat at an unchanged weight and therefore an unchanged BMI, which makes the number look better than the body. Cleveland Clinic excludes anyone over 65 from the standard chart for exactly this reason. Victoria's Better Health Channel describes 18.5 to 24.9 as the range for young and middle-aged adults and notes a suggested 22 to 26 for older Australians over 74, though it presents that as a researcher suggestion rather than guidance. This page shows an advisory from 65 and keeps the standard bands drawn, because shifting them without a clinical source behind it would be guesswork.
Where does the 703 in the pounds and inches formula come from?
It is a unit conversion folded into the formula, and it is rounded. A pound is exactly 0.45359237 kg and an inch is exactly 0.0254 m, so converting kg per m squared into lb per inch squared means dividing 0.45359237 by 0.0254 squared, which is 703.0696. Every chart prints 703. The difference is about one part in ten thousand, which cannot show at one decimal place but does mean the imperial shortcut and the metric formula disagree very slightly on the same body. This calculator converts feet, inches and pounds into metres and kilograms with the exact constants and then does one division, so both tabs return the identical number. The working box prints the 703 shortcut beside it so you can see the size of the gap.
Why is height squared in the formula when bodies are three dimensional?
Because squaring happened to fit the data Quetelet had, not because it follows from geometry. If people scaled up like solid objects, weight would grow with the cube of height and the exponent would be 3. Real people get proportionally less dense as they get taller, so the true exponent sits somewhere between 2 and 3. Work comparing the index against measured fat percentage points to about 2.5, which is the basis of Nick Trefethen's proposed alternative of 1.3 multiplied by weight in kilograms divided by height in metres to the power 2.5. The practical consequence of using 2 is that standard BMI runs slightly high for tall people and slightly low for short people, which is one more reason to read the band as a region rather than a verdict.