Wi-Fi Speed Test

How to tell Wi-Fi from broadband

One test cannot separate the two — but two tests can. Run this page over Wi-Fi, note the numbers, then connect an Ethernet cable (or stand directly next to the router with line of sight) and run the full speed test again.

The single most important rule: run the test while connected to the network type you want to evaluate. The page cannot tell which one you are on, so the only record of what was measured is what you were connected to at the time. Put the two runs in the same minute where you can, so peak-hour congestion does not explain the difference.

• Wired much faster than wireless: the Wi-Fi link is the limit. Nothing about your plan needs fixing.
• Both roughly equal and below your plan: the limit is further upstream — the router, the modem or the ISP.
• Both fine but calls still stutter: this is usually latency or jitter rather than bandwidth. Run the bufferbloat test.

The five changes that help most

1. Use the 5 GHz band. It has more usable channels and far less interference from neighbours. The 2.4 GHz band is for range, not speed.
2. Get closer, and get line of sight. Walls, floors and metal cost more throughput than any router feature can win back.
3. Raise the router off the floor. Radio waves spread better from a central, elevated position than from behind a television.
4. Reduce competing traffic. Background backups, cloud sync and video calls on other devices all share the same airtime.
5. Update the router firmware. Old firmware often means old channel-management and fewer security fixes.

Wi-Fi vs Ethernet — and Single vs Multi

This tool does not detect whether you are on Wi-Fi or Ethernet. No browser exposes that reliably, and a test that claimed to would be guessing. The measurement is the same either way — it measures the path from this device to the endpoint — so what changes the answer is which network you are actually connected to when you run it.

On the main speed test, the old “connection type” readout is replaced by a Connections control that does change the measurement: Single (one TCP stream) or Multi (four parallel streams). Use that when you want to see whether a lone stream or a filled path better matches your plan.

Run the test while connected to the network type you actually want to evaluate — that is the only way to know what you measured. Compare a Wi-Fi run with an Ethernet run from the same spot to isolate the wireless link.