Latency Test
Measure the round-trip time between your connection and 14 cloud regions across six continents. The test runs entirely in your browser, reports jitter alongside latency, and takes a few seconds.
Measures round-trip time to 14 cloud regions
Browsers cannot send ICMP packets, so this measures warm HTTPS round-trip time to each region and keeps the fastest of several samples. It reads roughly 1.5–2x higher than a command-line ping, but the relative differences between regions are accurate.
Reading your result
The number that matters most is the one at the top of the table: the closest region. That figure is the practical floor for anything you do online from this connection right now. No game server, video call or trading platform located further away can respond faster than the physical route allows.
The second thing worth reading is the spread. If your nearest region comes back at 25 ms and one on another continent at 260 ms, your connection is behaving normally — that gap is the speed of light through fibre plus routing overhead, and nothing on your end changes it. But if every region including the nearest one reads high, the delay is being added close to you: your Wi-Fi link, your router, or the first few hops inside your ISP's network.
Finally, check the jitter column against the latency column. Stable and slow is a geography problem. Fast but erratic is a local problem, and local problems are the ones you can usually fix.
Latency, ping, RTT and jitter — the actual differences
These four words get used as if they were interchangeable, and for most practical purposes they are. The distinctions still matter when you are trying to work out which number to trust.
Latency is the umbrella term for any network delay. Technically it can describe a one-way trip, though almost nothing measures that, because timing a single direction requires perfectly synchronised clocks at both ends.
Round-trip time is what is actually measured: out and back. It is roughly double the one-way delay, and it is the figure every game, every speed test and this page reports.
Ping began as the name of a diagnostic tool released in 1983 and became shorthand for the number it produces. "My ping is 40" means "my round-trip time is about 40 milliseconds."
Jitter is the only one of the four that measures consistency rather than speed. It is the average difference between consecutive samples, and it is frequently the real culprit when a connection "feels laggy" despite a respectable average.
What counts as good latency, by activity
There is no single threshold, because the tolerance depends entirely on how quickly the application needs to confirm your input with a server.
| Activity | Comfortable | Where it breaks down |
|---|---|---|
| Competitive FPS | Under 30 ms | Past 60 ms peeker's advantage becomes structural: the player swinging a corner sees you before you see them. |
| MOBA / RTS | Under 60 ms | Around 120 ms skill-shot leading and last-hit timing stop matching what your hands expect. |
| Fighting games | Under 25 ms | Frame-perfect inputs assume near-zero delay; every 16 ms is one dropped frame of advantage. |
| Video calls | Under 150 ms | Past roughly 300 ms round trip people start talking over each other because the natural conversational gap disappears. |
| Cloud gaming | Under 40 ms | Encoding and decoding already consume a latency budget, so the network portion has very little headroom. |
| Video streaming | Almost irrelevant | Players buffer several seconds ahead, so latency only affects how long the video takes to start. |
Why distance sets a floor you cannot negotiate
Light travels about 300,000 km per second in a vacuum, and roughly two thirds of that through glass fibre. London to New York is about 5,600 km, so a round trip covers 11,200 km and needs at least 56 milliseconds of pure travel time before a single router is involved. In practice the route is not a straight line and each hop adds processing, so 70-90 ms is a realistic best case.
This is why no router setting, gaming mode or premium plan will give a European player 20 ms to an American server. When people report that some product "halved their ping," what usually happened is that traffic was rerouted onto a shorter path, or that congestion and buffering were removed — not that physics was defeated.
The practical takeaway from the table above: pick the closest region your game or service offers, and treat everything else as a fixed cost.
When latency is your fault, and when it is not
Run the test twice, once on Wi-Fi and once on a cable, and compare the nearest region only. If the wired result is meaningfully lower or noticeably steadier, your wireless link is the bottleneck and it is worth fixing: change channel, move closer to the router, or switch to 5 GHz.
If both are similar and the nearest region still reads high, the delay is upstream. Common causes are an overloaded ISP uplink at peak hours, a congested local exchange, or a route that leaves your country before coming back. Those are worth raising with your provider, and a traceroute makes the case far more convincing than a single number.
And if the nearest region is fine but a specific game feels bad, the problem is the route to that game's servers rather than your connection in general. That is what the per-game tests are for.
Latency test questions
Is latency the same thing as ping?
In everyday use, yes. Strictly speaking latency is the general concept of network delay, round-trip time (RTT) is the specific measurement of a packet going out and the reply coming back, and ping is the tool that produces an RTT figure. When a game shows 'ping: 42 ms' and a monitoring dashboard shows 'latency: 42 ms', they are reporting the same round trip.
What is a good latency number?
Under 20 ms is competitive-grade and effectively invisible. 20-50 ms is comfortable for every genre. 50-100 ms is fine for most players but starts to matter in fighting games and tactical shooters. Above 150 ms you will feel a delay between your input and what happens on screen in any real-time game, though video streaming and web browsing stay unaffected.
Why is my latency high even though my internet is fast?
Bandwidth and latency are unrelated measurements. Bandwidth is how much data your line can move per second; latency is how long a single message takes to get a reply. A gigabit connection routed through a distant exchange can easily have worse latency than a modest connection with a nearby route. Distance, the number of network hops and congestion set latency; the size of your plan does not.
What does the jitter column mean?
Jitter is the average variation between consecutive samples. A route showing 40 ms latency with 2 ms jitter is stable and predictable. The same 40 ms with 25 ms jitter means individual packets are arriving anywhere between roughly 15 ms and 65 ms, which a game engine cannot smooth over — that is what produces sudden rubber-banding even though the average looks acceptable.
Why does this read higher than the ping command in my terminal?
The command-line ping tool sends ICMP packets, which the target machine bounces back at a very low level of the operating system. Browsers have no access to ICMP, so this test times small HTTPS requests instead. That adds TLS and browser processing overhead, so the number runs roughly 1.5-2x higher than raw ICMP. The comparison between regions is still accurate, because the same overhead applies to every one of them.
Does a wired connection lower latency compared to Wi-Fi?
Usually by a small but consistent amount. Ethernet typically removes 1-5 ms and, more importantly, removes most jitter. Wi-Fi adds a variable delay because the radio has to contend for airtime with every other device and nearby network. If your latency number looks fine but your jitter is high, the wireless link is the first thing to test.