Android Browser Benchmark Comparison: Firefox vs Chrome vs Opera vs AOSP Browser

BY Rajesh Pandey

Published 23 May 2013

chrome_opera_ff_browser

Over the last couple of weeks, all the popular browsers for Android – Chrome, Firefox and Opera – received major updates.

The Opera web browser based on the Webkit engine finally came out of beta earlier this week, while the Google team pushed out Chrome v27 that comes with the new full-screen mode, and stability and performance improvements. The latest version of Firefox also fixed some bugs and brought about some under-the-hood improvements.

With all the three most popular browsers on Android getting major updates, I decided to put them through some benchmarks to see how they perform. Keep in mind that technically Opera and Chrome are using the same rendering engine, while Firefox is using its own Gecko rendering engine. The AOSP browser is now defunct, and was last bundled by Google with the Android 4.1 update for the Galaxy Nexus and Nexus S. For comparison’s sake, I have also included the scores of the previous version of Chrome (v26), and Chrome Beta.

SunSpider Benchmark

sunspider_benchmark

Rankings -:

  1. AOSP Browser
  2. Firefox
  3. Opera
  4. Chrome Beta
  5. Chrome v27
  6. Chrome v26

The new version of Chrome brings about a major improvement in performance, but it is still nowhere near AOSP browser or Firefox. Except for the older version of Chrome, all the three other Webkit browsers performed more or less within ~3% of each other.

Firefox with its Gecko engine performed admirably and nabbed the second position. The AOSP Browser, which has not been updated in 10 months now, came out on top with a score of 1365ms.

Octane Benchmark

octane_benchmark

Rankings -:

  1. Chrome v27
  2. Chrome (Beta)
  3. Opera
  4. Chrome v26
  5. Firefox
  6. AOSP Browser

The results from the Octane benchmark paint a totally different compared to the SunSpider benchmarks. The outdated rendering engine of the AOSP browser shows its age here and the browser comes in last in the benchmark. The latest version of Chrome brought some major performance improvements, skyrocketing the browser to the top of the charts. For comparison sake, the Chrome v27 scored nearly a 1000 points more than the AOSP browser, which came in last.

Chrome Beta, Chrome v26 and Opera again performed very close to each other, which should not come as a surprise since they all are based on the same rendering engine. While Firefox performed admirably in the SunSpider benchmark, it came in second last in the Octane benchmark and scored very poorly compared to other Webkit browsers.

Conclusion

The above results are purely synthetic in nature and does not represent real world usage or performance. Personally, I prefer the stock AOSP browser. All the browsers, except for the AOSP browser struggle to render heavy websites like The Verge properly. There is a lot of lag involved when I try to quickly scroll up and down on such heavy websites. In fact, even on some light websites, all the browsers except for the AOSP one, would freeze for a moment or two. Even in the latest version of Chrome, that brought some noticeable performance improvement, scrolling is still not as smooth as it is in the AOSP browser. However, compared to the previous stable version of Chrome, there is a marked improvement.

If all the browsers performed the same, I would prefer to use Chrome because it automatically synchronises all my history, passwords and other data across my devices. While Opera and Firefox also provide similar functionality, I already use Chrome on my PC(s), which makes using Chrome on my Android device seamless and painless.

Keep in mind that all the browsers, except for the AOSP browser, use the same rendering engine as their desktop counterpart. This is probably a major reason why they feel so bloated compared to the AOSP browser. However, with Google and Opera shifting to Blink for Chrome and Webkit respectively, we should expect the performance and usability of these browsers to improve remarkably in future updates.