Browser Benchmark
Measure how fast your web browser runs real page work: JavaScript, DOM updates, Canvas 2D drawing, WebGL graphics, JSON and regular expressions. You get a score for each subtest and one overall score to compare browsers, settings or devices.
Last updated:
Close other heavy tabs for a more stable result. Nothing is uploaded.
This browser & device
Your previous runs
How to use the Browser Benchmark
- Close heavy tabs and apps and plug in your laptop, so other work does not slow the test.
- Pick a test length: Quick for a fast check, Standard for everyday use, Thorough for the most stable numbers.
- Press Start benchmark and leave the tab in the foreground until all six subtests finish.
- Read your scores: each bar is one subtest, and the big number is the overall score.
- Compare by running the test in another browser or with different settings; your previous runs are listed below.
What each subtest does
| Subtest | One operation | What it reflects |
|---|---|---|
| JavaScript | Prime sieve up to 20,000, sort 2,000 numbers, build and sum a 500-key object | JIT compiler and raw script speed: app logic, games, calculators |
| DOM | Create 100 elements, query by class, change a style, read layout sizes, remove them | How quickly pages update their content and recalculate layout |
| Canvas 2D | Draw 120 rectangles, circles and paths, a gradient and text on a 320 × 200 canvas | Charts, image editors, drawing apps and 2D games |
| WebGL | Render 6,000 triangles with a looping fragment shader, then read a pixel back | 3D graphics, maps and GPU-accelerated effects |
| JSON | Stringify and parse a 22 KB object with 150 records | Loading and saving data from web APIs |
| Regex | Find emails, rewrite dates, replace words and split a 12 KB text | Search, validation and text processing |
How the scores are calculated
Every subtest first runs briefly as a warm-up so the JavaScript engine can optimize the code. Then it repeats its operation in 50 ms slices for the chosen test length, pausing between slices so the page stays responsive. Only the time spent working is counted.
- Operations per second = completed operations ÷ working time in seconds.
- Subtest score = operations per second ÷ baseline × 1,000. The baselines are fixed constants chosen so that a modest 2-core computer scores roughly 1,000 per subtest; because they never change, scores from different runs, browsers and devices are directly comparable.
- Overall score = geometric mean of the subtest scores = (s1 × s2 × … × sn)1/n. A geometric mean treats a 2× gain in any subtest equally and stops a single very fast GPU result from hiding a slow DOM.
Example: subtest scores of 1,200, 900, 1,500, 2,000, 1,100 and 1,000 give an overall score of about 1,236.
Getting reliable results
- Keep the tab visible. Browsers throttle timers and rendering in background tabs, which lowers the score.
- Plug in your laptop and turn off battery saver; power-saving modes cap CPU and GPU speed.
- Close heavy tabs and apps such as video calls, games or large downloads.
- Repeat the test. Run it two or three times with the same length and compare the best or average score.
- Test extensions' impact by running once in a private window, where most extensions are disabled.
Frequently asked questions
What does the browser benchmark measure?
It runs six short workloads that resemble what web pages do: pure JavaScript computation (a prime sieve, sorting and object access), DOM work (creating, querying, styling and removing elements with layout), Canvas 2D drawing, WebGL rendering with a shader, JSON stringify and parse, and regular expression matching and replacing. Each workload is repeated for a fixed time and the number of completed operations per second is recorded.
How is the score calculated?
Each subtest score = operations per second ÷ a fixed baseline × 1,000, so doubling the speed doubles the score. The overall score is the geometric mean of the subtest scores, which means one unusually high or low subtest cannot dominate the result. If WebGL is not supported, it is left out of the overall score.
Why do my scores change between runs?
Browsers adjust CPU frequency, run garbage collection and share the processor with other tabs and apps, so results normally vary by a few percent. Use the Thorough length, close other programs, keep a laptop plugged in and run the test two or three times; compare the best or the average rather than a single run.
Which browser is fastest?
It depends on the device, browser version and workload, which is why this test reports separate subtest scores. Run it in each browser you use on the same computer, with the same test length, and compare the results in the history table. Browsers update their engines often, so re-test after major updates.
How is this different from the CPU benchmark?
A CPU benchmark focuses on raw processor speed, often across all cores. This browser benchmark measures how quickly the browser itself handles typical page work on the main thread, including rendering engines (DOM and Canvas) and the graphics pipeline (WebGL), so browser choice, extensions and settings affect it as well as hardware.
Can extensions or power saving lower my score?
Yes. Extensions that inspect or modify every page, battery saver modes, energy-saving browser settings and a background tab can all slow the test. Running the benchmark in a private window (where most extensions are disabled) is a quick way to see their effect.
Is any data uploaded?
No. All workloads run locally and the results, including your history of previous runs, stay in your browser. The GPU name shown comes from your browser's WebGL information and is not sent anywhere.
Are all the tools on FreeToolCabin tested for quality and accuracy?
Yes! Every tool on our site undergoes automated Q/A testing in real browser environments to ensure high precision, correct output, and zero errors. If you happen to spot a bug or have a suggestion, please contact us at [email protected] and our team will review it right away.