Measure your connection with real transfers
Prepare for a fair measurement
Connect to the network you want to test and pause large downloads, cloud backups and streaming if possible. Keep your device in its usual position when measuring Wi-Fi. The test automatically adjusts its transfer sizes to your connection. The test uses your real internet data allowance, so consider that before running it on a metered mobile connection. Press Start and keep this page open until the measurements finish.
Understand the four measurements
Download shows how quickly the browser receives test bytes, while Upload measures how quickly it sends them. Both are reported in megabits per second, or Mbps. Latency is the time for a small request to make a round trip. Jitter describes variation between consecutive latency measurements. These values are measured against the nearby test server serving this website. They describe this route and this moment, rather than guaranteeing performance to every website or application.
Compare results sensibly
Run another test at a different time if you suspect congestion. Compare the same device and network conditions when looking for improvements. A wired connection may behave differently from Wi-Fi, and a phone’s signal strength can change quickly. Short transfers are less precise on very fast connections because request setup and browser scheduling become a larger part of the measurement. Results do not measure your complete ISP service agreement, packet loss or maximum sustained throughput over a long session.
Stop or retry when necessary
Use Stop to cancel an active test. Completed measurements remain useful, while unfinished values are not replaced with guessed results. If the test endpoint is unavailable, the page reports the problem. Measurements are available when the test service is active. Test payloads are temporary synthetic bytes, not your personal files. A completed upload only returns its byte count. If your connection changes between tests, repeat under similar conditions before drawing a conclusion about the cause of a slowdown.