How the test works
Press Start. The browser tells the page every time it draws a frame, and the test counts those frames per second of visible time. Each complete second becomes one bar in the graph, so a run shows its history at a glance; the readout in the corner shows the current rate. The first half second after Start is discarded, since the frames right after a click are irregular, and time in a hidden tab does not count.
The load setting decides how much work the test adds to each frame. None draws only the graph, which is what an idle page costs. Medium draws six hundred moving dots in the panel on every frame, heavy four thousand. The dots are drawn with the same canvas calls a browser game uses, so the drop they cause is representative of animated pages, not just of this test. The run continues until you press Stop.
Under reduced motion the dots are drawn in place without moving; the drawing cost, and so the load, stays the same.
What the result means
The result is the average frames per second over the complete seconds of the run. The context line adds the worst and the best second, the frames the browser skipped, the longest single frame in milliseconds, the median frame rate of the run and the load.
- Average equal to the refresh rate, worst second within a frame or two. The machine keeps up. That is the normal result with no load on any modern device.
- Average close to the rate, a few skipped frames. Fine. Short hitches come from other programs, a notification, a background download, and do not show as stutter.
- Average well below the rate with no load. Something else is using the machine, or the browser is throttled: a power saver, hardware acceleration turned off, an extension. See the next section.
- Steady without load, a drop under medium load. Little headroom. Expect stutter in browser games and heavy animated pages; native apps are unaffected.
- A longest frame of several hundred milliseconds. A stall. Once is a garbage collection or a tab switch; regularly means the machine is swapping memory or overheating.
The number is capped by the screen. A 60 Hz monitor cannot show more than 60 fps in a browser, however fast the computer is; the refresh rate test tells you what that cap is.
What to do about low fps
Close other tabs, especially ones playing video, and quit programs that show activity in the task manager. Check that hardware acceleration is on in the browser settings; when it is off the processor draws every pixel and heavy load collapses. Plug in a laptop and set the power plan to balanced or performance, because most laptops slow the processor and graphics on battery. On a phone, turn off battery saver and let the device cool down; a warm phone lowers its clock speed and the fps with it. If one browser is much slower than another on the same page, the difference is the browser build, and updating it is the fix.
If the machine is old, the honest answer is that the heavy load is beyond it and nothing in settings will change that. The medium result is the one that predicts everyday pages.
When to return or claim the warranty
Low fps is almost never a warranty matter, because it is a performance characteristic rather than a defect. The exceptions are a new laptop or phone that cannot hold its own refresh rate with no load after a fresh start, or a device that overheats and throttles within minutes at medium load in a cool room. Record the result with the load and the refresh rate, note the room temperature and what else was running, and contact the seller within the return window; a device that cannot draw an idle page smoothly is not fit for purpose.
Frequently asked questions
What is a good fps result?
The same number as your refresh rate, with no dips. On a 60 Hz screen that is 60 fps, on a 144 Hz monitor 144. The dashed line in the graph marks the rate the test detected; bars that touch it mean the browser drew every frame. The number itself has no ceiling beyond the screen: a browser cannot draw more frames than the panel shows.
Why does my fps drop when I choose heavy load?
That is the point of the load. Heavy draws four thousand moving dots on every frame, which a desktop takes in its stride and a budget phone or an old laptop does not. A drop under load tells you how much headroom the machine has for games and animations in the browser; a drop with no load means something else is using the machine.
Does this measure my graphics card?
Only indirectly. Drawing in a web page goes through the browser, which uses the graphics chip for part of the work and the processor for the rest, so the result reflects the whole machine and the browser build together. Use it to compare browsers on the same machine or to see the effect of closing programs, not as a benchmark of the card alone.
Is fps the same as Hz?
Hz is how often the screen can show a new picture; fps is how many new pictures the browser actually produced. The refresh rate test measures the first, this test the second. On an idle machine they match. Under load fps falls below Hz, and that is when animation starts to stutter.
Why did the test stop counting when I switched tabs?
Browsers stop drawing a page that is not visible, so there is nothing to count. The run pauses and continues when you come back; the paused time is left out and listed in the result as a pause.