Internet Speed Test

Know exactly how fast your internet really is

You pay for a certain internet speed — but are you actually getting it? Buffering videos, laggy video calls, slow downloads and stuttering games are all signs of a connection that isn't performing the way it should.

An internet speed test measures the four numbers that matter — download, upload, ping and jitter — so you can see your connection's real capacity and responsiveness in seconds. It's the fastest way to confirm you're getting the speed you pay for, troubleshoot a slow connection, compare Wi-Fi to wired, or decide whether it's time to upgrade your plan.

This test runs entirely in your browser against a fast worldwide network, with no app to install and no sign-up. Press GO above, and within a few seconds you'll have a clear, honest picture of how your internet is really performing — plus a plain-English breakdown of exactly what it can comfortably handle.

Accurate testing, and how the numbers are measured

Like every internet speed test, this tool measures by transferring test data between your device and high-speed servers — here, Cloudflare's worldwide edge network, accessed directly from your browser. There's no account and no sign-up, and we don't store your results.

Tips for the most accurate reading

• Use a wired connection for your true line speed — Wi-Fi adds variability from distance, interference and your device's wireless limits.
• Close other apps and devices while testing, since anything else using the connection lowers your result.
• Test a few times, ideally at different times of day, as speeds vary with network congestion — an average is more reliable than one test.
• Compare Wi-Fi to wired — if wired hits your plan but Wi-Fi is slow, the bottleneck is your Wi-Fi, not your internet service.

If your wired speed is consistently well below your plan with nothing else running, that's worth raising with your provider — a record of repeated speed tests is useful evidence.

What a browser speed test can — and can't — tell you

Every speed test on the web, this one included, measures one specific thing: how much data your browser managed to move between your device and a set of test servers during a short window. That is a genuinely useful measurement, but it is not quite the same as "your internet speed", and being clear about the difference is what stops a single result from sending you down the wrong troubleshooting path.

Single-stream versus multi-stream throughput

This test opens several parallel connections rather than one, which is how modern speed tests reach the high figures people expect on fast lines. A single connection is limited by round-trip time and how much data can be in flight before an acknowledgement returns, so on a long-distance link one stream can top out well below your line's real capacity while four streams together saturate it. A multi-stream number reflects what your connection can do in aggregate — a fair proxy for several devices at once, or a download manager opening multiple connections — but pulling one large file from a single distant server may be noticeably slower, and that is normal rather than a fault.

Server distance and routing

Throughput and latency both depend on where the test data comes from and the path it takes. A nearby edge server flatters your connection; a distant one, or a congested route between your provider and the test network, drags the figure down for reasons unrelated to the line into your building. Ping is only partly a property of your connection at all — much of it is simply the physical distance the signal travels, plus a few milliseconds added at every router hop. This is why one fibre customer routinely sees a lower ping than an equally fast fibre customer testing against a server on another continent.

Wi-Fi, your device, and the browser itself

Testing over Wi-Fi measures the weakest link in a chain, and that link is usually the Wi-Fi rather than the internet service. Distance from the router, walls and floors, neighbouring networks crowding the same channel, an older 2.4 GHz-only adapter and other devices sharing airtime can all cap a result far below what the line delivers. Even on a wired connection the device can become the limit: a low-powered laptop, a saturated CPU, a security extension inspecting traffic or an active VPN will each shave something off. A browser is a convenient place to run a test, not the fastest possible one — on multi-gigabit connections, tools running closer to the operating system's network stack can report a higher ceiling.

What it genuinely does tell you

With all that acknowledged, a browser test still answers the questions most people actually have. Is this connection roughly in the range I pay for, or dramatically below it? Is upload far weaker than download? Is latency low enough for calls and gaming? Is the line steady, or does jitter spike? The four numbers above support those judgements well, provided you run the test more than once under consistent conditions. Treat any single result as one data point, not a verdict.

Why your number won't match your ISP's — or another speed test

Two tests run minutes apart on the same connection can differ noticeably, and neither is necessarily wrong — they measure slightly different journeys. A provider's own speed test often terminates on a server inside the provider's network, never crossing the peering point where congestion most often lives. That produces a flattering, technically accurate number that says little about how the connection behaves out on the wider internet. A third-party test crossing several networks answers a different and usually more relevant question.

Advertised plan speeds add to the confusion, because they describe the line rate rather than what an application ends up seeing. Protocol overhead means usable throughput always lands a little under the headline figure, so a connection sold as 100 Mbps testing in the low nineties is behaving exactly as expected. Congestion is the other variable: evening peaks in a shared neighbourhood move results for reasons unrelated to your equipment, which is why testing at a few different times of day tells you more than testing once.

Then there is the unit trap that catches almost everyone. Speed tests report megabits per second; browsers and download managers usually report megabytes per second — and there are eight bits in a byte. A 100 Mbps line downloads at roughly 12.5 MB/s at full tilt, so a transfer running at "12 MB/s" is not slow, it is saturating the connection. Our bandwidth converter switches between those units, the data storage converter handles file sizes when you are estimating transfer time, and the speed converter applies the same logic to physical units.

Finding the real bottleneck when the result looks low

A disappointing number is a starting point, not a diagnosis. Working through the chain in order isolates the cause far faster than changing several things at once:

• Run the test two or three more times. One low result can be a passing burst of traffic or a momentarily busy route; a repeated low result is a pattern worth investigating.
• Test wired, then over Wi-Fi. If the wired figure matches your plan and the Wi-Fi figure does not, the service is fine and the fix is in your wireless setup — router placement, channel, band, or an ageing adapter.
• Check what else is on the connection. A console downloading an update, a cloud backup or a second video call all consume capacity the test then cannot see.
• Try a different device. If one laptop is slow and everything else is fast, the problem is that laptop, not the line.
• Separate "slow internet" from "slow website". If throughput tests fine but particular sites crawl, the bottleneck is at the far end. Our Core Web Vitals score calculator is the better diagnostic there, since page weight and server response time govern that experience, not your Mbps.
• Keep a record before you call. A handful of timestamped wired results, taken with nothing else running, is far more persuasive to a support agent than describing the connection as sluggish.

If the local setup rather than the line needs attention, the IP subnet calculator handles addressing on a home or office network. And for anyone benchmarking their own responsiveness rather than the connection's, the typing speed test measures something different but works the same way — entirely in the browser, with nothing stored.

Frequently asked questions