The Complete Guide to Browser Benchmarks

The gloomy threat of massive energy blackouts highlights one thing: we depend on the internet. In our everyday lives, the internet is at our core. Whether they are apps, banking, TV, devices, car, or just browsing, we breathe it. Although browsing is just a small part of how the internet lives with us, it is still an important aspect.  

While browsing may seem like a small slice of our online experience, it’s the primary gateway to the web. Since the beginning of the www era, browsers have been the main modality for accessing and interacting with online content. Over time, they have evolved into sophisticated platforms, optimized for speed, security, and compatibility across countless devices and use cases. This evolution has resulted in a massive, diverse marketplace of browsers, each with unique technical specifications designed for different platforms and specific user needs. Today, we have a diverse landscape of browsers, each with its own technical specifications, performance characteristics, and target audiences.  

Users can access multiple competing browsers, such as Chrome, Firefox, Safari, Brave, and Comet. While some stick with the default browsers that come ready-installed on their device, others prefer to use custom browsers that offer more customization, privacy, or performance features.  

Regardless of preference, one thing is certain, we all rely on browsers. When choosing which browser to use, two key factors tend to guide our decision: 

• Usability 

• Popularity 

 Of course, the usability, the user experience, is often the most important factor when choosing a browser. But many times, we also let ourselves be guided by the popularity, or prevalence on the market, which is frequently shaped by performance metrics and technical insights that aren’t always visible to the average person. 

Image Source: Gemini AI

The performance of a browser is determined by the speed of computations, capabilities and the way it affects the overall speed of the system or device. On desktop systems, browsers have a minimal impact on performance. However, on mobile devices, the story is different. That is why most smartphone brands have their own dedicated browsers nowadays. They are specifically tailored so that they have minimal influence on the specific device’s performance. 

And this is where browser benchmarking comes into play. Just like any other benchmarks, a browser benchmark, such as Basemark Web 3.0 , evaluates the browser’s capabilities, independently of hardware or subjective user experience. 

Today’s browser market is like a crowded battlefield. With tech titans and independent developers constantly flooding the space with new solutions, choosing the right tool is a significant challenge for any end-user. This is why benchmarking is essential, it provides the objective data needed to select the optimal browser for any end-user. 

In short, before we go deeper into the topic of browser benchmarks, let’s clear out what a browser actually is.  

According to Browserstack – “a web browser is a software that enables users to access and view content on the World Wide Web. Its primary function is to locate and retrieve web pages, images, videos, documents, and other files from servers and display them on the user’s device.” 

A web browser is the essential link between the user and the internet. When you enter a URL into the address bar, the browser’s job is to locate and retrieve the appropriate files from a web server, then render and display the content for you in a readable and interactive format. 

Like other benchmarking software, browser benchmarks, such as Basemark Web 3.0, is designed to test the performance of browsers. Basemark Web 3.0, specifically, features a series of 20 tests grouped in 4 significant suites: Core, Graphics, Generic, Conformance.  

Image Text: Basemark Web 3.0

Once the tests are complete, you get a final score showing how well the browser performed. This score is then automatically posted on the Powerboard platform, where you can easily compare your browser’s results with other systems, hardware, or browsers.  

Launched as a free tool in 2016, Basemark Web 3.0 runs directly in any web browser and has been continuously updated and improved since its release. Thanks to its simple and straightforward interface, users can easily start a set of tests to measure the browser’s performance. 

When the benchmark starts, various tests will run sequentially in your browser. Since these tests are designed to push the browser to its limits, the visual output might look a bit chaotic. But don’t worry, these demanding visual and calculation routines are simply the test in progress. 

Next, let’s explore Basemark Web’s different testing suites. 

The Core Tests Suite has 3 computational tests, which evaluate the computing power through some of the most used algorithms in browser applications.  

• The Array test uses the Eratosthenes sieve for prime number determination. Since prime numbers do not have a straightforward mathematical law for determination, this test measures the raw computing power 

• The BitWise test has a total of 494 operations with integer numbers. It changes 5 predefined integers to binary representations, applies a mathematical operation and converts binary back to integer. The same operation is then performed with integers, and the results are compared 

• The Built-In Objects test evaluates the object and functions built-in capabilities of the browser. These functions and objects are widely used by all browsers, and they play a key-role in the functionality of the browser 

The Graphics Suite is the most demanding set of tests, as it evaluates the browser’s capability of interpreting and rendering graphics related content. Basemark Web 3.0 uses WebGL1.0.2 and WebGL 2.0 as graphics APIs to evaluate the browser performance. There are 7 tests included in this category: 

• Full scene rendering test with WebGL 1.0.2. A 3D scene is rendered in the browser, with focus on normal mapping, particle effects, anti-aliasing, depth of field etc. 

• Full scene rendering test with WebGL 2.0. Also, a 3D scene rendered in the browser, with additional focus on instance rendering and tile-based deferred lighting 

• Shader Pipeline test evaluates the browser’s performance when making basic shader calculations (vertex lighting, pixel lighting, bump mapping, parallax occlusion, fresnel etc.). 

• Draw-call stress test measures the framerate in relation to the amount of draw-call 

• Geometry stress test measures the framerate in relation to the amount of 3D polygons 

• Canvas test 

• SVG test 

The Generic Suite encompasses 6 tests with real-world use-case scenarios, revolving around the most used web frameworks and programming interfaces.  

• JQuery test – evaluates the use of jQuery as main JavaScript library for basic and common web operations 

• AngularJS test – evaluates the use of AngularJS as main JavaScript library 

• Backbone.js test – measures the performance when using Backbone.js as main JavaScript library 

• Crunch test – tests the speed of the browser against a large number of DOM and CSS style operations 

• DOM Search test – measures how quickly a browser can find and return elements on a webpage 

• DOM Create Source test – measures how fast the browser can create and insert new HTML elements into the DOM 

The Conformance Suite consists of 4 tests, which evaluate general browser capabilities: 

• CSS capabilities 

• HTML5 capabilities 

• Page load and responsiveness 

• Resizing capabilities 

To run the free Basemark Web 3.0 benchmark, the user only has to press the ‘Start’ button. No prerequisites are needed. Of course, to ensure the most accurate result, some conditions should be met, in order to alter the result as little as possible.  

Although these are not mandatory conditions for the benchmark to run, here are a few tips to ensure a proper and consistent evaluation:  

1. If you run the tests on mobile, it is recommended that you plug the charger. That is because portable devices are designed to minimize the stress on their hardware components when not charging. This way, the resource allocation is much more restricted, and that affects the results negatively. With less resources allocated, the completion of the stress tests becomes more difficult.
   
2. Keep the screen alive. If the screen turns off (as is the default case on mobile) the benchmark will be suspended. Then it will continue once the screen is turned on again. This will cause inconsistency in the testing process. The hardware components will cool down during the break, then they will start again “fresh”. 
3. Close other tabs. Any additional open tab inflicts more stress on the hardware and implicitly on the browser. For an accurate evaluation it is recommended to keep a minimum number of tabs open. 

After completing all the tests, the browser benchmark outputs a result, a score, where lower the score indicates a more badly performing browser. This final score includes all the individual scores for each test. It is calculated as a geometrical means of all the individual test results.  

Each individual test calculates either frames, or operations per second. In the end, being a performance test, it must measure the speed (*/s). The tests in the graphics suite evaluate the performance using the FPS (frames per second) while the other suites use “Number of Operations” per second.  

It is hard to say what is a good versus a bad score, since everything is relative to your system. But generally speaking, the lower the score, the worse is the browser performance.  

Image Text: Powerboard – System Comparison

When you run the benchmark, your results automatically go up on the Powerboard platform. The best way to understand your score is to check it against results from other people running the test on similar devices. That comparison will give you a clear idea of whether your browser is running as well as it should be. 

Image Text: An example of how the breakdown overview looks like in Powerboard after running the browser benchmark. 

By looking at the breakdown view of the results, you can determine how the browser performs under certain conditions, depending on the test type. For example, the Canvas tests can give clues about the browser’s responsiveness and performance when handling real-time 2D graphics and animations.  

In contrast to the synthetic graphics benchmarks, such as our GPU benchmarks, browser benchmarks paint a more accurate image about the behaviour of the browser. That is because browser benchmark tests are more specific and the situations they portray are met on a much wider scale. Therefore, we can say that a browser benchmark can give accurate data about how the browser behaves in different situations.   

Of course, for a clearer image, it is recommended that the benchmark should be run multiple times and use the average score from those results. Considering that Basemark Web 3.0 is a free tool, we encourage you to try it out (even with multiple different browsers) and see how the results vary from browser to browser, or how they scale against the already existing results in the Powerboard database. 

Part of our digital lives depends on our browser’s engine running smoothly. We have seen that in the ever-changing world of modern web technologies, simply relying on “a feeling” that the browser is loading fast isn’t the right strategy. To truly know if your browser is optimized for your tasks, you need hard, repeatable, objective data. This is what browser benchmarks do; they strip away subjective perception and eliminate the guesswork. The goal isn’t just to generate a high score, but to gain objective insight into how a browser handles crucial technical demands, from intense WebGL graphics to performance under common JavaScript frameworks.   

If you are ready to get that complete, objective picture, search for a browser benchmark, like Basemark Web 3.0, that utilizes an impartial methodology and covers a wide range of real-world scenarios. Use the objective data to choose the fastest browser for your device and use case, and ensure your time online is spent browsing fast, not waiting slow.