MB to GB Converter_
Both answers at once, because there are two and the difference is the reason you are here. A megabyte is 1,000,000 bytes to a disk manufacturer and 1,048,576 to Windows, so 1024 MB is either 1 GB or 1.024 GB depending on who is counting — and neither side is wrong.
The gap compounds: 2.4% at kilobytes, 4.9% at megabytes, 7.4% at gigabytes, 10% at terabytes. That single fact explains why a drive sold as 1 TB shows up as 931 GB, why your phone has less space than the box promised, and why two tools can report different sizes for the same file and both be telling the truth.
| Unit | Binary | Decimal |
|---|
- Input
- An amount and a unit, from bits up to petabytes, plus which standard the number you typed is already in. That last choice is the one other converters make for you silently.
- Output
- The converted amount under both standards side by side, a full table across every unit, and a sentence you can paste into a ticket.
- Processing
- Arithmetic in this tab. Powers of 1000 and powers of 1024, computed rather than looked up in a rounded table.
- Limits
- Beyond about 9 petabytes the arithmetic runs past the range where a double-precision number holds every integer, and results are reported in exponential form rather than pretending to exactness.
- Which standard should you use
- Storage vendors, network speeds and macOS use decimal. Windows, RAM and most programming use binary. If you are writing the number down, IEC symbols remove the ambiguity entirely — MiB and GiB can only mean 1024, which is why they exist. If you are reading somebody else’s number, assume binary for memory and decimal for anything sold by the gigabyte.
Why MB to GB has two answers
1000 and 1024, and how both became correct
Computers address memory in powers of two, so 1024 was the round number that mattered and engineers borrowed the SI prefix nearest to it. Kilo already meant 1000 everywhere else in measurement, and for small quantities nobody minded a 2.4% fudge. Then the numbers grew. At gigabytes the fudge is 7.4%, at terabytes it is 10%, and two industries had by then settled on different halves of the same word — storage marketing on the honest SI meaning, operating systems on the engineering one.
Where your missing disk space went
Nowhere. A drive advertised as 1 TB holds 1,000,000,000,000 bytes, exactly as promised. Windows divides that by 1024 three times to get gigabytes and displays 931. A 500 GB drive shows 466. A 2 TB drive shows 1,863. No space is lost, no partition is hidden, and formatting overhead accounts for a fraction of a percent rather than the 7%. The manufacturer and the operating system are using two different definitions of the same word, and only one of them puts it on the box.
KiB, MiB, GiB — the names that fix it
IEC published binary prefixes in 1998 precisely so that a number could be unambiguous in writing: a kibibyte is 1024 bytes and cannot be anything else, a mebibyte is 1,048,576, a gibibyte 1,073,741,824. Linux tools, some documentation and most standards bodies use them. Windows does not, and calls 1,073,741,824 bytes a gigabyte. If you are the one writing the number — in an API, a config file, a limit somebody else will implement against — the i removes an entire class of bug for the cost of one letter.
Bits are not bytes, and your connection is sold in bits
A 100 Mbps connection moves 100 million bits per second, which is 12.5 megabytes — the lower-case b is doing all the work in that abbreviation. This is why a download on a fast line runs at a tenth of the advertised number and nothing is wrong. Network speeds are quoted in bits and decimal, file sizes are usually reported in bytes and binary, and the two conventions differ by a factor of eight before the 1000-versus-1024 question even arises.
Which one your tools are using
Windows Explorer: binary, labelled KB and MB. macOS since Snow Leopard: decimal, labelled KB and MB. Linux ls -lh: binary by default, decimal with --si. df -h: binary. RAM: always binary, from the hardware up. Storage devices, network throughput and cloud pricing: decimal. When two numbers disagree by roughly 7%, this is almost always the reason — before you look for a leak, check the units.
Type the amount, read both answers
- 01Enter the amount and pick the units. Both standards are computed at once — there is no toggle deciding which answer you get.
- 02Set "Input is" to match where your number came from: binary for RAM and Windows, decimal for anything sold by the gigabyte.
- 03Read the table underneath for the same amount in every unit, so you can see where the two columns start to diverge.
- 04Copy the sentence if you need to put the answer in a ticket — it names the standard, which is the part that gets lost in a chat message.
A 1 TB drive that says 931 GB
Nothing is missing. The manufacturer counted in powers of 1000 and the operating system is dividing by 1024, three times over.
1 TB = 1,000,000,000,000 bytes
931.32 GB (the same bytes, divided by 1024 three times)
An upload limit written in the wrong unit
A limit documented as 100 MB is implemented as 104,857,600 bytes by one team and 100,000,000 by another. Files between the two sizes fail unpredictably.
max_upload = 100 MB
max_upload = 104857600 # 100 MiB # or 100000000 for 100 MB decimal
RAM that does not match the invoice
Memory is binary all the way down — the hardware is addressed in powers of two. A 4096 MB module is 4 GiB, and there is no decimal reading of it that means anything.
4096 MB installed
4 GiB exactly (4,294,967,296 bytes)
A download slower than the connection
The line is sold in bits and the download is reported in bytes. Divide by eight before deciding anything is wrong.
100 Mbps
12.5 MB/s
Every unit, both ways
| Unit | Decimal (SI) | Binary (IEC) | Gap |
|---|---|---|---|
| Kilobyte | 1,000 bytes — KB | 1,024 bytes — KiB | 2.4% |
| Megabyte | 1,000,000 — MB | 1,048,576 — MiB | 4.9% |
| Gigabyte | 1,000,000,000 — GB | 1,073,741,824 — GiB | 7.4% |
| Terabyte | 1,000,000,000,000 — TB | 1,099,511,627,776 — TiB | 10.0% |
| Petabyte | 1,000,000,000,000,000 — PB | 1,125,899,906,842,624 — PiB | 12.6% |
Every percentage in this column is computed in the test suite from the two powers rather than typed in, because a gap that grows with each step is exactly the kind of number that gets copied wrong once and repeated forever.
Habits that prevent a unit argument
- Write MiB and GiB when you mean 1024. One letter removes the ambiguity permanently, and every standards body already uses them.
- Store limits in bytes, not in units. A config saying
104857600cannot be misread; one saying100 MBis read two ways by two teams. - Suspect the units before you suspect a leak. A 7% discrepancy between two size readings is almost always 1000 against 1024, not missing data.
- Divide by eight when the number came from a network. Connections are sold in bits and files are measured in bytes, and that factor dwarfs the 1000-versus-1024 gap.
- Check what your own tooling reports.
ls -lhanddf -hare binary;ls -lh --siis decimal; cloud consoles are almost always decimal because that is what they bill on.
Where the two definitions cause real damage
A limit that means two things
An upload cap documented as "10 MB" is 10,000,000 bytes to one implementer and 10,485,760 to another. Files in between are accepted by one service and rejected by the next, which is a bug nobody can reproduce.
Capacity planning that quietly under-provisions
Ordering storage in decimal terabytes and planning in binary ones costs 10% of what you thought you had. At scale that is a rack, and it is discovered when the volume fills.
Bits mistaken for bytes in a budget
Bandwidth is quoted in bits per second and transfer is billed in bytes. Reading one as the other is wrong by a factor of eight in whichever direction hurts most.
Rounded tables copied between documents
A table that says 1 GB is "about a billion bytes" is fine in prose and wrong in code. Compute the value rather than transcribing a rounded one; the errors compound with every unit you climb.
Standards, ranges and precision
- Decimal (SI)
- Powers of 1000. kB, MB, GB, TB, PB. Used by storage manufacturers, network throughput, cloud billing, and macOS since 10.6.
- Binary (IEC)
- Powers of 1024. KiB, MiB, GiB, TiB, PiB, standardised in 1998. Used by Windows Explorer, RAM, and most programming interfaces.
- Bits
- Eight bits to the byte, handled as its own conversion rather than as another power — it is not one, and folding it into the exponent produces silently wrong answers.
- Precision
- Computed in double-precision floating point. Past roughly 9 petabytes that stops holding every integer, and results switch to exponential form rather than implying an exactness they do not have.
- Network
- None from tool code. A test sweep calls every function this page uses with
fetchandXMLHttpRequestreplaced by stubs that throw, so a stray request fails the build instead of shipping. Disconnect from the network and the page still works.
Questions about MB, GB, and the two standards
How many MB are in a GB?
Either 1000 or 1024, and which one depends on who is counting. Storage manufacturers, network providers and macOS use 1000. Windows, RAM and most programming use 1024. Both answers appear above at once, because picking one silently is what makes the two numbers impossible to reconcile.
Is 1024 MB equal to 1 GB?
It is exactly 1 GiB — one gibibyte — and 1.024 GB in decimal. If you are looking at RAM or at Windows, 1024 MB is your 1 GB. If you are looking at a drive specification or a cloud bill, 1 GB is 1000 MB and 1024 MB is slightly more than that.
Why does my 1 TB drive show only 931 GB?
Because it holds 1,000,000,000,000 bytes exactly as advertised, and Windows divides by 1024 three times to reach gigabytes rather than by 1000. Nothing is missing and nothing is hidden — formatting overhead is a fraction of a percent, while the gap here is 7.4%. A 500 GB drive shows 466 for the same reason.
What is the difference between MB and MiB?
A megabyte in the SI sense is 1,000,000 bytes; a mebibyte is 1,048,576. The IEC published the binary prefixes in 1998 so that written numbers could be unambiguous, and the extra letter is the whole mechanism. If you are specifying a limit that somebody else will implement, MiB says what you mean and MB does not.
How do I convert MB to GB?
Divide by 1024 if the number came from Windows, RAM or a programming interface, and by 1000 if it came from a storage specification, a network figure or a cloud console. The converter above does both at once so you do not have to decide before you see the answers.
Why is my download slower than my connection speed?
Because the connection is sold in bits and the download is reported in bytes. A 100 Mbps line moves 12.5 megabytes per second at best — the lower-case b is the entire difference. That factor of eight is much larger than the 1000-versus-1024 question and is usually what people are actually noticing.
Which standard should I use in my own code?
Store and compare raw byte counts, and only convert for display. Where you must name a unit, use the IEC form when you mean 1024 — a config that says 104857600 or 100 MiB cannot be misread, and one that says 100 MB will eventually be implemented two ways by two teams.
Is MB bigger than GB?
No. A gigabyte is a thousand or 1024 megabytes depending on the standard, so GB is the larger unit in both. The order from smallest is byte, kilobyte, megabyte, gigabyte, terabyte, petabyte.
Does anything I type here get sent anywhere?
No. The work is JavaScript running in this tab. Every function it calls is covered by a test that stubs fetch and XMLHttpRequest to throw, so a request that slipped in would break the build rather than reach a server — and you can confirm it for yourself by disconnecting and carrying on.