Speed test says 500, but downloads crawl at 60? Most of the gap is megabits vs megabytes. Here's the math and the other causes worth checking.
Your speed test says 500 and your game launcher shows 60. The most common reason is not a fault: speed tests report megabits per second (Mbps), while download managers, browsers and game launchers show megabytes per second (MB/s). A byte is eight bits, so divide the speed test number by 8 and you get the best case you can see in a download window. If the gap is bigger than that, the cause is almost always somewhere other than your line: the server, the number of connections, your disk, your Wi-Fi, or a VPN.
Key Takeaways
- Speed tests show megabits per second (Mbps); download windows show megabytes per second (MB/s). Divide Mbps by 8 for the ceiling in MB/s.
- A 100 Mbps line tops out near 12.5 MB/s; a 500 Mbps line near 62.5 MB/s; a 1 Gbps line near 125 MB/s.
- After the unit conversion, the usual culprits are a slow or distant server, a single TCP connection, slow disk writes, weak Wi-Fi, and a VPN.
- A speed test measures your path to one test server. A download measures your path to a different server that you do not control.
- Compare like with like: wired, one download from a fast host, and the speed test number divided by 8.
Why does a 500 Mbps plan download at 60 MB/s?
Because the two numbers use different units. Network speeds are quoted in bits because that is how links are rated. File sizes and download progress are quoted in bytes. One byte is eight bits, so the conversion is just division:
| Speed test result | Best-case download |
|---|---|
| 50 Mbps | about 6.25 MB/s |
| 100 Mbps | about 12.5 MB/s |
| 300 Mbps | about 37.5 MB/s |
| 500 Mbps | about 62.5 MB/s |
| 1000 Mbps (1 Gbps) | about 125 MB/s |
So a 500 Mbps plan that downloads at 60 MB/s is performing almost perfectly. The connection is fine; the advertised figure is simply in a different unit. For the full picture of bandwidth, latency and jitter, see internet speed metrics explained.
One more wrinkle: some tools show MiB/s (mebibytes, 1,048,576 bytes) while others show MB/s (megabytes, 1,000,000 bytes). The difference is about five percent, so it rarely matters, but it can explain a small leftover gap.
What if the gap is bigger than 8x?
Once you have done the division and still see much less, the cause is usually one of the following. They are listed roughly from most to least common.
The remote server is the bottleneck
A speed test deliberately talks to a nearby, well-connected server that exists to be fast. A download comes from whoever hosts the file: a game publisher's CDN node, a software mirror, a small website. If that server is busy, rate-limited per user, or far away, it will send you less than your line could take. No change on your side fixes this. A quick check is to download something large from a different, big host and see whether the speed changes. If it does, the line is fine and the first server was the limit.
One connection versus several
Many speed tests open several parallel TCP connections and add them up; others use a single stream. A single TCP connection is limited by round-trip time and by how fast the sender is willing to grow its window, so on a fast but distant path one connection often cannot fill the pipe. M-Lab describes its NDT test as a single-stream measurement of a connection's capacity for bulk transport, which is why such tests can read lower than multi-stream ones. A plain browser download is a single stream too, while a download manager or game launcher that splits a file into several parts can go faster. How the two approaches differ is covered in speed test methodology.
Latency to the server
The farther away the server, the longer each round trip, and the slower a single connection ramps up. This is physics, not a fault; see latency vs distance for how much delay to expect on a given route.
Protocol overhead
Every packet carries headers, and encryption and retransmissions add more. With standard 1,500-byte packets, TCP/IP headers and Ethernet framing alone take roughly 5 to 6 percent of the line rate. The file data you actually save is always somewhat less than the raw line rate. The effect is real but modest, and it varies by protocol and path, so treat it as a small reason the real number sits a little under the ceiling, not as something to chase.
Your disk cannot write that fast
At gigabit speeds the network may be quicker than the storage. A nearly full drive, a slow external disk, an SD card or USB stick, or antivirus scanning every incoming file can cap the speed. If the speed starts high then drops to a steady low number, or stalls while the disk activity light is busy, suspect the disk.
Wi-Fi link rate is not throughput
The "1200 Mbps" shown in your Wi-Fi settings is the radio link rate, not the data you can actually move. Real throughput is typically well below it, and it falls with distance, walls, interference and other devices sharing the channel. Test with a cable to rule this out. The same factors are discussed in why speed test results vary.
A VPN is in the path
A VPN adds encryption work and sends your traffic through another server, so both speed and latency change. If a VPN is on during the download but you tested without it (or the reverse), the numbers will not match. Why a VPN slows your connection explains the four causes.
Other traffic and queues
If someone else is streaming, uploading or backing up to the cloud, your download shares the line. Heavy upload traffic can also slow downloads and make everything laggy, because queues fill up; bufferbloat explained covers that, and upload vs download speed explains why the upload side matters more than its size suggests.
How to compare fairly
- Convert units. Divide the speed test's Mbps by 8 to get MB/s.
- Use a cable for both tests, and pause other downloads and cloud sync.
- Turn the VPN off or on for both, not for one of them.
- Download a large file from a fast, well-known host and compare to the converted number.
- If a particular service is slow but others are fast, the cause is that server or the route to it, not your connection.
You can run BrowserInsight's network speed test to get your baseline. It reports download and upload in Mbps, so divide by 8 before you compare it with a download window.
Frequently Asked Questions
Is Mbps the same as MB/s?
No. Mbps is megabits per second and MB/s is megabytes per second. One byte is eight bits, so 8 Mbps equals 1 MB/s. The capitalisation matters: lowercase "b" is bits, uppercase "B" is bytes.
My speed test shows 100 Mbps and my download is 11 MB/s. Is something wrong?
No. 100 Mbps divided by 8 is 12.5 MB/s, and 11 MB/s is about 88 percent of that, which is normal once protocol overhead and server conditions are counted.
Why is Steam or another game launcher slower than my speed test?
After the unit conversion, the usual reasons are the download server's own load or per-user limits, the distance to it, and disk write speed. Some launchers also let you choose a download region, which is worth trying.
Why is my download fast at first and then slow?
A common cause is a cache or buffer filling up. The first burst goes into memory, then speed falls to what the disk or the server can sustain. Server-side rate limiting after the first few seconds is another cause.
Does a faster plan fix slow downloads?
Only if your line really was the limit. If the server, Wi-Fi, disk or VPN is the bottleneck, paying for more bandwidth changes nothing.
Conclusion
The first thing to do is divide by 8. In many cases that closes most or all of the gap. If it does not, the bottleneck is usually the server you are downloading from, the number of connections, your disk or Wi-Fi, or a VPN, and each of those is quick to check on its own.
