Convert MB to MiB

A megabyte is a million bytes. Storage manufacturers have always used this decimal meaning, which is why their capacities look larger than what a computer reports.

A mebibyte is 1,024 kibibytes, or 1,048,576 bytes. Linux tools and memory sizes generally mean this even when they print "MB".

The factor is 0.953674, and almost nobody carries that around. Rounded to 0.95 it is off by 0.39 % — which stays invisible on small numbers and turns into a whole unit somewhere around 1,000 MB.

That is the number worth knowing before you round: not the error itself, but where it stops being ignorable. Below that point the shorter factor is the sensible one; above it, use the field above, which never rounds until it prints.

One MiB is 1,024 of the unit below it; one MB is 1,000. On this page that is the difference between 1.0486 MB and 1 MB — 4.9 % — and the gap grows at every step up the scale, which is why it is a rounding error on a photograph and a visible chunk of a hard disk.

This is the whole of the missing-storage mystery, and on this page it is worth 4.9 %. A drive sold in MB holds exactly what the label says; Windows divides by 1,024 instead of 1,000, keeps the decimal name, and reports 1 MB where the box said 1.0486. macOS has counted these in the decimal units since 10.6, which is why the same drive can look two sizes on two machines — nothing is missing and nobody is rounding, the same bytes have two names.

A file manager shows 24.8 MB. The form says maximum 25 MB. The upload is refused, twice, and nothing on either screen explains it. What happened is that the two numbers are in different units wearing the same abbreviation: the file manager divided by 1,048,576 and the server divided by 1,000,000, so 24.8 on one side is 26.0 on the other.

This is the single most common way the binary-decimal split costs somebody real time, and it is entirely invisible while it is happening. There is no error message that says "your megabyte is not our megabyte". The file fails, the obvious next move is to compress or crop it, and the person doing it never learns that the margin they needed was one per cent rather than ten.

Limits written by providers are usually decimal, because they are set by people writing specifications rather than by an operating system rendering a directory listing. A cap of 25 MB is 25,000,000 bytes, a "max 10 MB" field is 10,000,000, and a 2 MB avatar limit is 2,000,000. When a limit is enforced in code it is almost always an integer of bytes compared directly, and that integer was written by hand as a round decimal figure.

The size you can see, by contrast, comes from whatever is displaying it. Windows counts in mebibytes and writes MB. Most Linux file managers do the same. The macOS Finder counts in megabytes and means it. So the two figures in front of you are typically measured differently, and the difference always runs the same direction: the file is larger than it looks.

File Explorer has used 1,024-based steps since long before the IEC coined the mebibyte in 1998, and it has never adopted the newer names. A 26,004,685-byte file is shown as 24.8 MB, which is a correct division by 1,048,576 and an incorrect label. Microsoft has kept the label for compatibility with three decades of documentation, screenshots and habit.

The way out is the properties dialog, which prints the exact byte count underneath the friendly one. That number is unambiguous and is what a server compares against. If a limit matters — an upload, a mailbox quota, a submission portal — read the byte count and compare integers rather than reading two rounded figures in two conventions and hoping.

Snow Leopard changed the Finder to report storage in decimal units, so a drive sold as 500 GB finally displayed as 500 GB and files became slightly larger numbers overnight. The change removed one confusion and created another: the same file now reads differently on a Mac and on a PC, with both screens saying MB.

That is worth knowing when a size is quoted to you by somebody else. "It’s 26 megabytes" from a Mac user and "it’s 24.8 megabytes" from a Windows user can be the same file, and the discrepancy is not a transfer problem, a compression difference or a corrupted copy. If two people disagree about a size by roughly five per cent, the unit is the first thing to check.

The useful arithmetic runs the other way from the one this page is named for. If a limit is 25 MB decimal, the largest file that fits is 23.84 MiB, which is what your file manager will show as 23.8 MB. For a 10 MB limit the target is 9.5 as displayed; for 50 MB it is 47.7; for 2 MB it is 1.9. The rule is to multiply the stated limit by 0.954 and treat the result as the number on your own screen.

A rounder version is easier to remember and safer: subtract five per cent from the stated limit and aim below that. The lost margin is trivial — half a mebibyte on a 25 MB cap — and it absorbs both the unit gap and whatever the upload wraps around the file. A file that fits with room to spare uploads once.

Email is the case where the unit gap is the smaller problem. Attachments are base64-encoded so that binary data survives a text transport, and base64 turns every three bytes into four characters — a 33 per cent increase before the line breaks are counted. Where the stated limit applies to the assembled message, a 25 MB ceiling accepts a file of roughly 18 MB, and no unit conversion explains the missing seven.

Web forms behave differently and better. A browser upload uses multipart form data, which carries the bytes unmodified with only a few hundred bytes of headers and boundary markers around them. So a 10 MB limit on a web form really is about the file, and the only correction needed is the decimal-versus-binary one. Knowing which of the two situations you are in tells you whether to trim five per cent or a quarter.

A gap of half a per cent would round away and a gap of fifty would be obvious. This one sits in the band where it changes nothing about how a file feels and everything about whether it clears a threshold. Photographs, documents and short videos all land in the range where a stated limit and a displayed size differ by less than a rounding step you would question.

It also compounds upward, which is why the same misunderstanding costs more the larger the numbers get. The binary and decimal units diverge by 2.4 per cent at kilo, 4.86 at mega, 7.37 at giga and 9.95 at tera, because each step multiplies the discrepancy by 1.024 again. On a mail attachment that is one file rejected; on a storage array it is the better part of a disk.

Once the unit question is settled, the remaining gap is real and the fix depends on what the file is. Photographs and video are already compressed, so putting them in a ZIP saves almost nothing — a per cent or two, occasionally less than the archive overhead. Reducing the dimensions is what actually works: halving the pixel width of an image removes about three quarters of the data, and it is usually invisible on a screen the recipient will view it on.

Documents behave differently. A PDF is often large because of the images inside it rather than the text, so re-exporting at a lower image resolution can take a great deal off without touching legibility. Office files, logs and data exports compress well and a ZIP is worth trying first. And where the file is only slightly over — which after this conversion it usually is — trimming a single embedded image or an unused page is faster than re-encoding everything.