Internet Speed Test

Measure your connection in three quick phases: latency first, then download, then upload. Everything runs in your browser against this site's server, and no result is stored.

Run the test

The test transfers about 25 MB of download data and 5 MB of upload data. On a metered or mobile connection, that usage counts towards your allowance.

Ready when you are. Press Start test to begin.

No phase running yet.

Ping (average)—milliseconds
Ping (minimum)—milliseconds
Jitter—milliseconds
Download—Mbps
Upload—Mbps

What this speed test actually measures

A speed test does not measure “the internet” in the abstract. It measures the path between your device and one specific server, at one specific moment, using one specific method. This test runs three phases in order. First, it checks latency by sending ten small requests to the server and timing each round trip. Second, it downloads a 25-megabyte response as fast as your connection allows and converts the transfer time into megabits per second. Third, it uploads a 5-megabyte block of data generated in your browser and times that transfer too.

Bandwidth is not the same as latency

Bandwidth, usually quoted in megabits per second (Mbps), is how much data can move per second once a transfer is underway. Latency, measured in milliseconds as ping, is how long a single small request takes to travel to the server and back. Think of a motorway: bandwidth is how many lanes it has, while latency is how long the journey takes. A video stream cares mostly about bandwidth, because it moves large amounts of data continuously. An online game or a video call cares more about latency and jitter, because late or unevenly timed packets cannot be fixed by having extra lanes. Jitter is the variation between successive ping times; this page calculates it as the average absolute difference between consecutive pings. Low, steady latency usually feels more responsive than a slightly lower average with wild swings.

Why your results change from run to run

Several bottlenecks can limit a test before your broadband line itself does. The server's location matters: a test server in another country adds distance, extra routing hops, and often slower paths, so the same connection scores differently against different servers. Your home network matters just as much. Wi-Fi signal strength, interference from neighbouring networks, an old router, or a housemate streaming in 4K can all cap the result below what the line delivers. Time of day matters too, as networks are busiest in the evening. That is why a single test proves little; run the test a few times, at different times, and look at the pattern rather than one headline number.

How to get a result you can trust

For the fairest measurement, connect your computer to the router with an Ethernet cable, pause downloads, streams, cloud backups, and game updates on every device, and close other browser tabs that might be busy. Stand close to the router if you must use Wi-Fi, and prefer the 5 GHz network if your router offers one. Run the test two or three times and compare the median result with the speed your provider advertises; note that advertised figures are usually “up to” values, and a wired result consistently far below them is worth raising with your provider.

Reading the numbers

As rough guides, one high-definition video stream needs about 5 Mbps, a 4K stream about 25 Mbps, and a large household sharing all of these at once benefits from 100 Mbps or more. Upload speed matters for video calls, sending large files, and backing up photos; many cable and DSL plans upload far more slowly than they download, which is normal rather than a fault. Ping under 30 ms is excellent for gaming, under 75 ms is comfortable for most uses, and jitter in single figures means a stable connection.

Frequently asked questions

How much data does this test use?

About 30 MB in total: roughly 25 MB downloaded and 5 MB uploaded, plus ten tiny ping requests. Avoid repeated runs on a tightly capped mobile plan.

Why is my result lower than the speed I pay for?

Usually the limit is Wi-Fi, a busy home network, a distant test server, or peak-time congestion rather than the line itself. Test again with a cable, at a quiet time, before concluding the line is faulty.

Why is upload so much slower than download?

Most consumer cable and DSL products are deliberately asymmetric, because households historically downloaded far more than they uploaded. Fibre-to-the-premises plans are more often symmetrical.

Does this test store or share my results?

No. The measurements are calculated in your browser from the timing of the transfers, and the figures are not saved or sent anywhere other than the test requests themselves.

Related tools