Four numbers, four different stories
Download speed describes how quickly your browser receives data. Upload speed describes how quickly it sends data. Both use megabits per second (Mbps), while a file size is often measured in megabytes (MB). Eight bits make one byte: at a constant 40 Mbps, a 100 MB file would take about 20 seconds before overhead and changes in speed.
Latency is the response delay to the measurement endpoint when this test is idle. Jitter describes variation between consecutive response delays. A connection can have high bandwidth and still feel inconsistent when response times fluctuate. This page does not measure packet loss, latency under load or promise a gaming or streaming quality score.
What the bunny and dial tell you
During download, the bunny moves down toward its lower carrot. During upload, it climbs the ladder toward the upper carrot. It pauses to nibble before repeating the movement while that phase remains active, then rests when you stop or finish. Its movement illustrates transfer direction; it is not a network measurement or a progress indicator. The needle follows the current download or upload estimate on a labelled logarithmic scale from 0 to 1,000 Mbps. Larger results remain visible as numbers even when the needle reaches the end.
You can pause the animation while keeping the test running. Your device’s reduced-motion setting also stops the decorative movement and needle transitions. Numeric results remain available either way.
Why two tests can disagree
This tool measures the route from your browser to Cloudflare’s edge network at that moment. Wi-Fi interference, a VPN, other devices, browser activity and the chosen sample sizes can change results. Your ISP plan is not a guaranteed speed to every destination.
Light mode schedules up to 20.4 MB of download and upload payload combined; Extended schedules up to 144.4 MB. Adaptive stopping can use less, while retries and protocol overhead can use more. Larger samples can give fast links more time to settle. Both modes are short snapshots, not certified line-capacity measurements.
A useful next step
Repeat only when you need a comparison: first near your router, then in the room where the connection feels slow. Keep the test mode and background activity similar. For another everyday comparison, the percentage change calculator can quantify the difference between two download results. A reduction in latency is usually desirable, so state the direction along with the percentage.
The measurement engine is Cloudflare’s open-source speedtest module, used here with our own interface and test sizes. QuickCalculators is an independent website. Engine license