Mbps vs MB/s: Reading Your Real Data Transfer Rate

I spent an embarrassing amount of time once arguing with my ISP. I had just upgraded to a 300 Mbps fibre plan, but my download manager stubbornly showed numbers like 36 MB/s. My first thought was that I was being throttled. My second thought, after some math, was that I had been confusing two units that look almost identical but differ by a factor of 8. The plan was fine. My reading of it was wrong.

That single factor of 8 is the source of nearly every "why is my internet slow" complaint that turns out to be no problem at all. Once you internalize it, bandwidth stops feeling like a black box.

The one rule that explains everything: 8 bits = 1 byte

A bit is the smallest unit of digital information, a single 0 or 1. A byte is a group of 8 bits. That is the whole trick. Everything downstream follows from it.

Internet service providers advertise speed in bits per second: Kbps, Mbps, Gbps. The lowercase b means bits. Your file manager, browser, and operating system report transfer speed in bytes per second: KB/s, MB/s, GB/s. The uppercase B means bytes.

Because one byte holds 8 bits, you divide the bit-rate by 8 to get the byte-rate:

100 Mbps = 12.5 MB/s (100 ÷ 8)
300 Mbps = 37.5 MB/s
1 Gbps = 125 MB/s (1000 Mbps ÷ 8)

So when my 300 Mbps line delivered 37.5 MB/s, it was running at exactly its theoretical ceiling. Nobody was throttling anything. The plan was quoted in bits; the download was reported in bytes. You can check any plan against its real-world byte ceiling with the data transfer rate converter, which shows every unit side by side from one input.

Why ISPs use bits and your computer uses bytes

This split is not a conspiracy, though the marketing angle is real. Networks measure throughput in bits because data physically travels the wire one bit at a time. Bits are the native engineering unit of a connection, the way horsepower is native to engines. So Mbps is what network gear has always reported.

There is also a presentation bonus. "100 Mbps" looks faster than "12.5 MB/s," even though they describe the same line. A bigger headline number sells better, so the industry never had a reason to switch to bytes.

Storage went the other way. Files, disks, and memory are counted in bytes because a byte is the smallest addressable chunk a program works with. A 1 GB file means roughly one billion bytes, not bits. That mismatch — connection in bits, file in bytes — is exactly what makes download estimates feel wrong until you reconcile the units.

Stepping between Kbps, Mbps, Gbps, and Tbps

Within the bit family, networking uses decimal (SI) prefixes that step by 1000:

1 Mbps = 1000 Kbps
1 Gbps = 1000 Mbps = 1,000,000 Kbps
1 Tbps = 1000 Gbps

So a 1500 Kbps video stream is 1.5 Mbps, and a 2.5 Gbps link is 2500 Mbps. No 1024 anywhere — that is the second trap. The 1024-based step belongs only to binary units like Kibit and Mibit, which you will rarely meet outside specific spec sheets. If you accidentally apply 1024 to ordinary Mbps, you inflate every figure by about 2.4%, and the error compounds as you scale up to gigabit and beyond.

A clean way to remember it: decimal prefixes (kilo, mega, giga) on bit-rates always mean powers of 1000 in networking. Reserve the "kibi/mebi" mental model for storage and the explicit binary toggle.

Worked example: converting a 500 Mbps line to MB/s

Say a colleague gets a 500 Mbps business line and wants to know the real download speed.

Start with the bit-rate. 500 Mbps.
Divide by 8 to reach bytes. 500 ÷ 8 = 62.5 MB/s. That is the byte-per-second ceiling your file manager would show on a perfect transfer.
Sanity-check against a file. Downloading a 5 GB game update: 5 GB is 5000 MB, and 5000 ÷ 62.5 = 80 seconds at full speed. Note that 5 GB is also 40,000 megabits (5000 MB × 8), and 40,000 ÷ 500 = 80 seconds — same answer, whichever unit you work in.
Account for reality. Real transfers rarely hit 100% because of protocol overhead, server limits, Wi-Fi, and congestion. Expect 80–95% of the ceiling, so plan for roughly 53–60 MB/s in practice.

The headline figure — 62.5 MB/s — is the number that matters, and it comes straight from dividing the advertised speed by 8.

Putting it to work

The bits-versus-bytes gap shows up everywhere once you start looking:

Sizing video streams. A 4K stream needs about 25 Mbps, which is 3.125 MB/s. On a shared 100 Mbps office line (12.5 MB/s), you can sustain three or four simultaneous streams before the link saturates.
Reading vendor spec sheets. A NAS quotes sequential throughput in MB/s while the switch it connects to is rated in Gbps. Convert 1 Gbps to 125 MB/s and line it up against the drive's MB/s figure to find the actual bottleneck.
Settling arguments. When someone insists their 1 Gbps fibre should pull a 1 GB file in one second, the conversion does the talking: 1 Gbps is 125 MB/s, so a 1 GB file takes about 8 seconds at full speed, not one.

The three mistakes worth avoiding

Treating Mbps and MB/s as the same unit. They are 8x apart. A 200 Mbps plan is 25 MB/s, not 200 MB/s. Always ask whether a number is in bits (lowercase b) or bytes (uppercase B) before you compare two figures.

Mixing decimal and binary prefixes. Networking rates step by 1000. Using 1024 quietly inflates everything and the drift grows with scale.

Confusing a rate with a size. Mbps is a speed (per second); MB is an amount. A 1 GB file over a 100 Mbps line is not instant — 1 GB is 8000 megabits, so it needs about 80 seconds at full speed. When you want the time directly rather than the rate, pair the rate converter with a download time calculator that folds file size and speed into one estimate.

The takeaway

You do not need a chart taped to your wall, just one fact: ISPs sell bits, your computer counts bytes, and the gap between them is exactly 8. Divide the advertised Mbps by 8 and you have the MB/s you will actually see. Once that clicks, every plan, spec sheet, and "is my internet broken" moment becomes a quick mental division instead of a mystery.

Made by Toolora · Updated 2026-06-13