KB, MB, MiB, and GiB: File Size Explained

KB, MB, MiB, and GiB are labels for quantities of digital information, but they do not all use the same conversion step. Decimal units such as MB use powers of 1,000; binary units such as MiB use powers of 1,024. Knowing which one a website or device means can explain why a file appears to be over a limit even when another app shows a smaller number.

Start with bytes

A byte is the basic unit used to count a file’s stored data. Many interfaces display a rounded value such as 24 MB, but the exact file size is an integer number of bytes. When a limit matters, the exact byte count and the service’s definition are more reliable than a rounded display.

The letter case also matters:

• b commonly represents a bit.
• B represents a byte.
• One byte contains 8 bits.

That difference is especially important when comparing a file’s size with an internet speed. A file might be 100 MB, while the connection is 100 Mb/s. The values use different units and cannot be compared directly without conversion.

Decimal and binary units

Decimal prefixes use powers of 1,000. Binary prefixes use powers of 1,024. The IEC binary prefixes add an i to distinguish them clearly:

| Unit | Bytes | Typical interpretation | |---|---:|---| | 1 kB | 1,000 | Decimal kilobyte | | 1 MB | 1,000,000 | Decimal megabyte | | 1 GB | 1,000,000,000 | Decimal gigabyte | | 1 KiB | 1,024 | Binary kibibyte | | 1 MiB | 1,048,576 | Binary mebibyte | | 1 GiB | 1,073,741,824 | Binary gibibyte |

In strict technical notation, the decimal symbol is kB with a lowercase k; many interfaces and everyday writing use KB. Check the product’s own labeling when precision matters. The important distinction is whether the conversion uses 1,000 or 1,024.

At small sizes the gap is modest. It grows with each unit. One GiB is 73,741,824 bytes larger than one GB, so a file reported as 1 GiB is about 7.4% larger than 1 GB. A system that compares bytes against a strict 1 GB cap may reject a file near 1 GiB even if a file manager rounds both to “1 GB.”

Convert a file size yourself

To convert bytes to decimal megabytes, divide by 1,000,000:

bytes ÷ 1,000,000 = MB

To convert bytes to mebibytes, divide by 1,048,576:

bytes ÷ 1,048,576 = MiB

For example, a file of 25,000,000 bytes is 25 MB in decimal units. In binary units, it is about 23.84 MiB because 25,000,000 ÷ 1,048,576 ≈ 23.84. The file did not shrink; only the unit used to describe it changed.

To convert in the opposite direction, multiply. A 25 MiB limit equals 25 × 1,048,576 = 26,214,400 bytes. A 25 MB limit equals 25 × 1,000,000 = 25,000,000 bytes. If a service provides a maximum in bytes, compare directly to that number.

The Data Storage Converter can convert common storage units. When you use a converter, check whether it distinguishes decimal from binary units; a dropdown that says only “MB” may not answer what a specific service means.

Why two apps show different sizes

Different applications can use different unit conventions, round to different decimal places, or show allocated storage instead of the exact logical file length. A file manager may display a rounded size; a browser’s upload message may show a different unit; a cloud service may compare the exact bytes to a fixed maximum. These numbers can all describe the same file.

Filesystems can also allocate data in blocks. The amount of disk space consumed may differ from the file’s logical byte length, particularly for very small files. For a web upload limit, the service generally cares about the uploaded content size and may also account for request overhead or related metadata. Use the service’s stated method when it provides one.

A display such as “25 MB” is not necessarily an exact promise that the file is below a 25 MB limit. It may be rounded from a value slightly above the threshold. If the upload is rejected, inspect the exact bytes or create a safer margin below the limit.

When comparing two files, compare like with like. An operating system may show size on disk, while a website reports the uploaded file’s content length. A ZIP archive is a different file from the uncompressed folder it contains, and its size depends on how compressible the contents are. A video or JPEG may already use compression, so placing it in another archive may barely change the total. Write down the exact byte count and where it came from if a discrepancy is causing a failed submission.

Read upload limits carefully

Before resizing or compressing a file, find out what the recipient means by its maximum:

1. Is the limit per file or for the total submission?
2. Does the site specify MB, MiB, or an exact byte count?
3. Does it include only the file, or also a group of attachments?
4. Does the service report a maximum or a target recommendation?
5. Is there an additional restriction on format, dimensions, duration, or page count?

If the instructions only say “10 MB,” do not assume the unit convention. Check the help page or test with a file comfortably below the apparent limit. Avoid trimming a file to the rounded boundary; a result a little smaller is easier to validate.

For image uploads, size in bytes is separate from pixel dimensions. A 1 MB image could have a large width and height if compressed efficiently, while a smaller-dimension image could still be large because it is saved in a less compressed format. A service can enforce both rules. Check the actual requirement rather than treating “under 2 MB” as a complete upload specification.

PDFs have another pair of properties to keep distinct: file size and page size. A 2 MB file might contain one high-resolution scan or many mostly textual pages. Page dimensions such as A4 or Letter describe the physical layout, not how many bytes the file occupies. If a document fails an upload check, confirm whether the issue is bytes, number of pages, or page dimensions before changing it.

For storage planning, add byte counts before converting to larger units. If a folder contains 400 files averaging 2 MB each, a rough total is 800 MB in decimal units, but the actual sum may differ because rounded displays hide variation. For an exact archive or backup, use the completed package’s reported size rather than multiplying a rounded average.

When sharing a size in a ticket or email, state both the display and the exact byte count if the discrepancy matters: “The file is 26,214,400 bytes (25 MiB).” That statement is less ambiguous than “about 25 MB.” If you are comparing a folder with an upload limit, first create the archive if the recipient expects one, then check the archive’s size. The total of the uncompressed files does not predict the archive size exactly because different file types compress by different amounts.

Use the same convention throughout a calculation. Do not convert a decimal MB figure to MiB using 1,000,000 as the divisor and then call the result binary. Keep the unit beside each intermediate number; this small habit catches many “my file should be below the limit” mistakes before you repeatedly compress or resize the wrong asset.

If the recipient’s instructions are vague, ask for the maximum in bytes or check its official help page. A screenshot of the rejected upload can also show whether the system reports a file-size threshold or another validation error. Confirm the actual error before modifying a document that may already meet the byte limit.

File size is not transfer speed

Storage units describe an amount of data. Transfer-rate units describe how quickly data moves. A rate such as 20 Mb/s means 20 million bits per second if the prefix is decimal. To estimate time for a file size in bytes, convert bytes to bits by multiplying by 8, then divide by bits per second. Real transfers also depend on throughput, overhead, and server behavior.

For example, a decimal 100 MB file contains 800 megabits. At a sustained 20 Mb/s, its ideal transfer time is 40 seconds. If the display reports 2.5 MB/s, that is equivalent to 20 Mb/s before network overhead. The Data Transfer Rate Converter can help compare rates, while a separate time estimate requires the file size and an actual or assumed throughput.

Avoid unit-related surprises

• Compare exact bytes when you are close to an upload cap.
• Keep b and B distinct when reading network speeds and file sizes.
• Check whether an app uses decimal or binary prefixes.
• Do not infer image dimensions from its file size or vice versa.
• Leave a margin below a service’s maximum if the service’s rounding is unclear.
• When reporting a size to someone else, include the unit and, if necessary, the byte count.

Frequently asked questions

Is 1 MB the same as 1 MiB?

No. One decimal MB is 1,000,000 bytes. One MiB is 1,048,576 bytes. The difference comes from decimal versus binary conversion.

Why does my computer say a file is smaller than the website says?

They may use different unit conventions or rounding. Compare the exact byte count with the service’s limit when available. The underlying file may not have changed.

Is KB or KiB correct?

They mean different quantities. Strictly, kB is 1,000 bytes, while KiB is 1,024 bytes. Some software uses KB informally for binary quantities, so check its documentation if the distinction affects a limit.

How many bytes are in 1 GB?

In decimal units, 1 GB is 1,000,000,000 bytes. In binary units, 1 GiB is 1,073,741,824 bytes. Some interfaces label binary sizes with GB even though they mean GiB.

Does compressing a file change its dimensions?

Not necessarily. Compression can reduce the number of bytes while leaving image dimensions or document page dimensions unchanged. Some export methods also resize or rasterize content, so inspect both file size and output properties.

Why does a drive advertise more capacity than my computer shows?

Manufacturers commonly state capacity using decimal units, while some operating-system displays use binary quantities with decimal-looking labels. Formatting and reserved system space can also affect available capacity. Compare bytes or clearly labeled GB and GiB before concluding that storage is missing.

Use the unit the recipient uses

When a file is near an upload limit, check the exact bytes and confirm the service’s unit convention. Use decimal MB for decimal limits and MiB for binary limits; do not rely on rounded labels alone. Once the size is safely within the limit, verify the file still meets any separate format, resolution, or content requirements.