Internet Speed Test

About this test

What's actually being measured

This test uses Cloudflare's @cloudflare/speedtest JavaScript library — the same one that powers speed.cloudflare.com. It performs HTTP downloads and uploads of various sizes against Cloudflare's edge network, then uses the browser's PerformanceResourceTiming API to extract precise timing data.

Why these specific numbers

• Download / Upload (Mbps): peak sustained throughput. The test runs multiple parallel streams at different file sizes to find the real ceiling, not just a single-stream measurement.
• Latency (ms): idle round-trip time to the nearest Cloudflare edge — what your ping would show. Anything under 50 ms is excellent, 50–100 ms is fine for everything except competitive gaming, 100–200 ms is noticeable, 200+ is uncomfortable.
• Jitter (ms): variance in latency. High jitter wrecks VoIP and video calls even if your bandwidth and average latency look fine. Sub-5 ms is great, 10–20 ms is the threshold where you start hearing it on a phone call.

What can throw off the result

• WiFi vs Ethernet: WiFi adds latency, jitter, and a hard ceiling well below your ISP's plan. To know your real connection, plug in.
• Background traffic: a Steam download, cloud backup, or someone streaming Netflix in the next room steals bandwidth.
• VPN / WARP: adds 5–30 ms latency and usually caps throughput at 50–500 Mbps depending on the service.
• Browser overhead: JavaScript-based tests are bounded by the browser's network stack. On gigabit+ connections the test can underreport by 5–15%. For real wire-speed measurements use iperf3 against a private endpoint.
• Local network bottleneck: a 100 Mbps cable, an old switch, or weak WiFi caps the result regardless of your ISP plan.

What the verdict means

The "Tank's verdict" line at the end is just a vibe-check based on the numbers — it doesn't compare you to other users or report anything anywhere. Tank is offering an opinion. He has many opinions.