Unit Converter
All units
- nmnanometer
10,000,000- µmmicrometer
10,000- mmmillimeter
10- cmcentimeter
1- mmeter
0.01- kmkilometer
0.00001- ininch
0.393700787402- ftfoot
0.0328083989501- ydyard
0.0109361329834- mimile
0.00000621371192237- nminautical mile
0.00000539956803456
Nothing you enter here is stored, logged or shared.
Convert between metric and imperial units for length, weight, temperature, area, volume, speed, data and time. Every conversion uses the exact defined factor, so results are not thrown off by floating-point error.
How to use
- Pick a category, then the units to convert from and to.
- Type a value; the result and every other unit update as you type.
- Use the swap button to reverse the direction.
Where the factors come from
Most metric-to-imperial conversions are exact, which surprises people who expect them to be approximations. Since the international yard and pound agreement of 1959, an inch is defined as exactly 25.4 millimetres and a pound as exactly 0.45359237 kilograms. Those are definitions, not measurements, so the conversion carries no error at all.
The metric side is anchored in the SI base units, which since 2019 are all defined from physical constants rather than artefacts - the kilogram from the Planck constant, the metre from the speed of light. What that means in practice is that every factor in the length, mass and volume categories here is a rational number known exactly.
The exceptions are the units that were never defined precisely in the first place, and calendar time is the main one. A month has no exact length, so the figure used is one twelfth of a Julian year of 365.25 days, about 30.44 days. Treat any month or year conversion as an average.
Exact and rounded
Because the factors are exact, the arithmetic can be too. Conversions here are computed as rational numbers and rounded once, at the very end, so an inch comes back as 2.54 centimetres rather than 2.5399999999999996.
That artefact is not a bug in the tools that show it, it is what happens when 0.0254 is stored as a binary floating-point number, since it has no exact binary representation any more than a third has an exact decimal one. The error is tiny and usually harmless. It becomes visible when it is displayed, and it accumulates when conversions are chained - going metres to feet to inches to millimetres through floating point drifts a little at each step.
Each result is labelled as exact or rounded so you can tell which you are looking at. Temperature is the one category where the operation itself is different: the scales have different zero points, so a conversion needs an offset as well as a factor. This is why 20 degrees Celsius is not twice as warm as 10, and why ratios of temperatures are meaningless unless you are working in kelvin.
The units that are not what you think
Liquid measure is the big one. A US gallon is 3.785 litres, an imperial gallon 4.546 - about 20 per cent larger. Pints differ both in size and in composition: an imperial pint is 20 fluid ounces, a US pint is 16, and the fluid ounces themselves are different sizes. Both systems are listed separately here rather than guessing which you meant. Recipes are a common casualty, particularly older British and American cookbooks.
Mass has its own traps. A troy ounce, used for precious metals, is about 31.1 grams against the ordinary avoirdupois ounce's 28.35. Tons come in three: the US short ton of 2,000 pounds, the imperial long ton of 2,240, and the metric tonne of 1,000 kilograms, all of which are sometimes written just ton.
In length, the US survey foot was very slightly longer than the international foot and was officially retired at the end of 2022, but it persists in older land records where the difference matters across long distances. A nautical mile is 1,852 metres exactly, unrelated to the statute mile, and a knot is one of those per hour.
Decimal and binary data units
A kilobyte is 1,000 bytes and a kibibyte is 1,024, and the reason your drive appears smaller than the box said is that the two conventions are both in use on the same machine.
Storage manufacturers use decimal prefixes, so a 1 TB drive holds 1,000,000,000,000 bytes and is honestly labelled. Operating systems have historically reported in binary units while writing them with decimal names, so the same drive shows as 931 GB where it means 931 GiB. Nothing is missing; the space is being counted in units of 1,024. macOS switched to decimal reporting years ago, which is why the same drive reads differently on a Mac and on Windows.
The IEC prefixes - KiB, MiB, GiB, TiB - exist to say which you mean, and both sets are available here. Two habits are worth keeping: memory is always binary, since address lines make powers of two natural, while network throughput is decimal and measured in bits rather than bytes, so a 100 Mbps link moves at most about 12.5 MB per second before overhead.
What the 2019 redefinition of the SI changed
Until 2019 the kilogram was a lump of platinum-iridium in a vault near Paris, and every other kilogram in the world was defined by comparison with it. The problem was that it drifted: copies distributed to national institutes diverged from the original by tens of micrograms over a century, and there was no way to say which had changed.
The redefinition fixed the Planck constant at an exact value and derived the kilogram from it, so the unit is now reproducible by anyone with the right apparatus rather than by comparison with an artefact. The metre had already been defined from the speed of light since 1983, and the second from a caesium transition since 1967.
For everyday conversion none of this changes a number. It matters because it makes the whole system self-consistent and exact by definition, which is the reason the conversion factors here are rational numbers known precisely rather than measurements with error bars.
Significant figures and honest rounding
A conversion cannot add precision that the input did not have. If a length was measured as 5 metres to the nearest metre, reporting 16.404199475065617 feet implies a measurement a hundred million times more precise than the one you took. Carrying the digits through the arithmetic is correct; presenting them all is not.
The convention in scientific work is to keep the number of significant figures the input had, and to do the rounding once at the end rather than at each step. Rounding early compounds: a value rounded, converted, rounded and converted again can drift well beyond the tolerance anyone intended.
Which digits to keep is a judgement about the measurement, not about the maths, which is why the results here are computed exactly and labelled as exact or rounded rather than truncated to a fixed number of places. You keep the digits your input justifies.
When unit errors have gone badly wrong
The Mars Climate Orbiter was lost in 1999 because one team supplied impulse data in pound-force seconds while the navigation software expected newton-seconds. The spacecraft approached the planet far lower than intended and broke up. The cost was around 125 million dollars, and the failure was not a calculation error but an interface where the unit was assumed rather than stated.
In 1983 an Air Canada 767 ran out of fuel at altitude because the load was computed in pounds during the airline's transition to metric, giving the aircraft roughly half the fuel it needed. It was landed on a disused runway with no engines. Nobody died, largely through the crew's skill and luck.
The lesson both incidents teach is the same and it is cheap to apply: put the unit in the variable name, the column name, the API field and the label. timeout_ms rather than timeout; weight_kg rather than weight. A unit that lives only in a comment or in someone's memory is a unit that eventually gets assumed wrongly.
Temperature behaves differently from everything else
Every other conversion here is a multiplication. Temperature needs an offset as well, because the scales have different zero points, and that single fact produces most temperature errors. Twenty degrees Celsius is not twice as warm as ten; the ratio is meaningless unless you work in kelvin, where zero is absolute.
The distinction that trips people up is between a temperature and a temperature difference. A reading of 20 °C is 68 °F, but a rise of 20 °C is a rise of 36 °F, because a difference is scaled without the offset. Anything describing a change - a tolerance, a gradient, a rate per hour - uses the second rule.
Absolute zero is a real floor: -273.15 °C, 0 K, -459.67 °F. A conversion producing a value below it is an error rather than a very cold result, which is why this converter refuses it rather than returning a number.
The units where marketing and engineering disagree
A drive labelled 1 TB holds 1,000,000,000,000 bytes, which is honest decimal usage. Windows reports it as 931 GB because it counts in units of 1,024 while using the decimal name, so the same drive appears to lose 7% of its capacity between the box and the operating system. Nothing is missing; two conventions are in use at once. macOS switched to decimal reporting years ago, which is why the same drive reads differently on each.
Network speeds add a second confusion by being measured in bits rather than bytes. A 100 Mbps connection transfers at most about 12.5 MB per second before protocol overhead, and a gigabit line tops out around 118 MB per second in practice. The lowercase b is the whole difference, and it is a factor of eight.
Memory is always binary, because address lines make powers of two natural, so a 16 GB module really is 16 GiB. The IEC prefixes - KiB, MiB, GiB - exist to say which convention you mean, and using them in documentation removes an argument that otherwise recurs forever.
Frequently asked questions
- Why do other converters return 2.5399999999999996 instead of 2.54?
- Because they do the arithmetic in floating point, which cannot represent 0.0254 exactly. This tool works in exact fractions and rounds once, at the end, so one inch comes back as 2.54 cm. Each result is labelled exact or rounded, so you always know which you are looking at.
- Is a US gallon the same as a UK gallon?
- No, and the gap is large: a US gallon is 3.785 liters while an imperial gallon is 4.546, about 20 percent more. Pints and fluid ounces differ too - an imperial pint is 20 fluid ounces, a US pint is 16. Both sets are listed separately here rather than picking one and hoping you meant that one.
- Why does my 1 TB drive show as 931 GB?
- Because the two figures count differently. Drive manufacturers use decimal prefixes, so 1 TB is 1,000,000,000,000 bytes. Your operating system reports the same space in binary units of 1024, properly written GiB. 1 TB is 931.32 GiB - nothing is missing, it is the same space labelled two ways. Both units are available here.
- Why is temperature handled differently from length?
- Because the scales do not share a zero. Converting length is a single multiplication, but 0 degrees Celsius is 273.15 kelvin, so temperature needs an offset as well as a scale factor. That is why 20 degrees Celsius is not twice as warm as 10, and why a converter that only multiplies gets Fahrenheit wrong.
- How long is a month or a year here?
- A year is the Julian year of 365.25 days and a month is one twelfth of that, about 30.44 days. Neither has a single correct definition, since calendar months run from 28 to 31 days, so treat any month or year conversion as an average rather than an exact answer.
- Is anything I enter sent to a server?
- No. The conversion runs in your browser and nothing you type is stored, logged or shared.
Last updated:
Include the steps, expected result, and actual result. Remove secrets and personal data.