Samsung Galaxy S3 gets benchmarked

by: Ankit BanerjeeMay 5, 2012


The Samsung Galaxy S3 was officially announced two days back. By now, heart rates have begun to slowly go down and the world is catching its breathe again. Not helped by the overwhelming hype in the build-up to the final release, there have been some mixed reactions to the device, with the biggest gripe being the lack of the (hoped for) 12MP camera. Despite the disappointment of some users, no one can deny that, hyped or not, the Samsung Galaxy S3 is one of the best Android devices available at the moment.

With its impressive specifications, we all expected Samsung’s latest flagship device to be powerful. Initial benchmark test results, conducted by Anandtech,  certainly point towards the Galaxy S3 being one of the most powerful smartphones available now. Let’s take a look at some of these results. All images are courtesy of Anandtech.

Sunspider Javascript Benchmark

The Sunspider Benchmark is a Javascript benchmark that tests the core JS language only, and not other browser APIs. It compares different versions of the same browser and different browsers to each other. Lower is better.


BrowserMark is another benchmark test that measures browser performance on mobile devices, by measuring device performance in Javascript and HTML rendering. Higher is better.

As you can see from the results above, the Samsung Galaxy S3 significantly outperforms every other high-end device available. The large difference is likely because of the browser code and the Android 4.0.4 OS. Therefore, the performance boost is owed more to the firmware, rather the Exynos 4 Quad processor.


Vellamo is an Android benchmark test developed by the Qualcomm Innovation Center that tests the important aspects of mobile web experience including rendering, Javascript, user experience, and networking. Higher is better.

With the exception of the US version of the HTC One X powered by Qualcomm’s own Snapdragon S4 processor, The Galaxy S3 again outperforms competition by a large margin.


The GLBenchmark tests the quality and performance of the OpenGL ES implementation, and covers gaming, navigation, and user experience. Higher is better.

The onscreen GLBenchmark results are v-sync (vertical synchronization) limited, i.e. the performance matches the refresh rate of the screen (if an image is refreshed 60 times a second, the game won’t perform above 60FPS) and therefore the benchmark results in this case are inconclusive. But in the off-screen tests, the Galaxy S3 outperforms the competition again, including the iPhone 4S, which was the previous leader in the GPU arena.

Basemark ES 2.0

The Basemark ES 2.0 benchmark is another test used for checking graphics performance of mobile hardware and is specifically designed to to benchmark OpenGL ES 2.0 hardware. Higher is better.

Both benchmark tests above are v-sync limited as well, but the Samsung Galaxy S3 still manages to pull slightly ahead in the Taiji test.


As expected, the Samsung Galaxy S3 is one of the best Android devices available at the moment, and the long wait for the device was certainly worth it.

What are your thoughts? Are you surprised at the exceptional Galaxy S3 benchmark tests results? Let us know in the comments section below.

  • how is the battery life! does it drain like the Galaxy S2?

    • AppleFUD

      Samsung has had pretty good battery life with their devices of late. The Galaxy Note, from what I’ve read, has great battery life with a 5.5″ 720P screen and it uses a 25,000 battery where the SIII is using 2,1000. Considering the SoC in the SIII should be more efficient and the screens have the same resolution, though smaller on the SIII, I would expect it to have pretty good battery life.

      • Ummm….

        Dude if the S3 had a 21,000 Mah battery it would last for like 10 days without charging and the battery wound be the size of a brick. You mean 2,500 Mah and 2,100 Mah.

        • AppleFUD

          got an extra zero in there LOL. . . but the comma was in the right place.

    • I guess , its the same as the S-II

    • Davide

      My S2 on Sprint lasts a REALLY long time. Almost 2 days with medium usage, and 1 day with heavier usage. Samsung uses good batteries.

      The only company that has a horrible reputation with batteries is HTC. I’d stay away if battery life is a concern.

  • AppleFUD

    Samsung’s last Exynos was a leader and now it has twice as many cores and the gpu speed has been bumped, thus no surprise it is again the leader. However, this will be the year of the A15 chips and when they start showing up at the end of summer we are going to see some impressive bench marks.

    Nonetheless, the whole problem I have with the SIII, other than it still looks like the 3GS — hardware buttons?!?!?! seriously?!?!?! — is that it will be once again fubared by the US carriers and totally dependent on them for updates. . . and since I don’t need all the game playing power, web would be nice, I’ll be sticking with the unlocked Nexus line.

    • Kookas

      Eh.. dunno what you’re seeing. It looks nothing like the 3GS. The S2 and some of Samsung’s other Galaxy phones did look like the spitting image of the iPhone 4S, but the S3 (and the similarly shaped Nexus) is far too curved and basically looks completely different.

    • Striderevil

      Hmmm….. HTC One X in the US is running on a dual core S4 snapdragon processor which is a A15 chip. Dual A15 cannot beat a Quad A9 as an A15 is 40% more better than an A9. So if one A9 gives 100% and A15 gives 140%, 4 x A9 = 400% while 2 x A15= 280% lol.

      • AppleFUD

        Don’t disagree with your numbers and the S4 is doing better than some quad-cores (not all SoC are created equal)

        Example: OMAP 5 Vs Quad Core A9!

        thus I really only look toward OMAP 5 & Exynos, clearly TI & Samsung are able to get more out of their chips than others. The Exynos has consistently kicked butt and the OAMP 4 isn’t bad at all — just use the PlayBook and play a 1080P video and continue to multitask, no other tablet can pull that off with a dual-core 1GHz from a year ago. So the OMAP 5 & Exynoas A15 dual core should prove to be better than we think when compared to what else is available and hopefully they will have quad-core versions soon after, not to mention the overall efficiency gained from going to A15.

      • Danny

        The one S4 is a9 NOT a-15. Its based of the a-15 architecture but its not the a-15. Therefore its acutely suprising that the dual a9s are competing and sometimes out performing some of the quadcores. Also the I’m pretty sure the dual core a15s will out perfrom all quadcore a9s. Look at the OMAP series.

        • Knnamdikalu

          Ur an idiot go to qualcoms own website and u will see it there. There a15 cores…

    • Terrence Greene

      what are you surprised about… samsung has always put true hardware buttons on their phones.

  • EviLRome0

    i imagine the comparo is for the internaitonal s2. how does it compare with the us/ca version with the 1.5ghz S3 processor?

    • Striderevil

      hmm… slower. The S3 is twice as slow as the S4 to give you a Snapdragon very own benchmark comparison and they built it.

  • bau

    Are these tests results influenced by the number of the pixels available for each device? Because different screen resolutions have different total pixels amount:
    – a resolution of 960 x 640 = 614.000 total pixels, while
    – a resolution of 1280 x 720 = 921.600 total pixels

  • Whooo

    So à galaxy s2 is the better or same as a htc one X…Who make this benchmark…Samsung ?

    • Veronica

      These are well known neutral benchmarks, and the article provided almost a dozen of them.

      If the htc one fails to be tangibly better than a year old Galaxy S2 on all of the tests, blame htc for making a poor product.

  • Stains

    ThEM grAPhs must BE wRonG, Iphone Win them ALL!!!1one!11!!! GalaXY S three TImes aS DUMB mire LIEK it. YOU anDUMBroid fanbois dON’t know A quaLIty pHOne if it BIT your dick OfF. YOU oul PrObaBle like it anyWAY since iT wOuld be ThE closest you WouLD ever HAVE been to Getting hEaD.

    • Yonatan Horwitz

      That was a joke, right?
      You can’t possibly be that retarded… including the misspellings and all…

    • ringsgeek

      Wow. You’re really smart.

  • Jaranjames

    Come on android authority give the people what they want. We want to see a real comparison the Sprint Version HTC EVO 4G LTE vs the Samsung Galaxy S3. And for shits and giggles the benchmark of the quadcore (which will not be a US version) vs the dual core. Thank you Sprint for taking something and Finally making it better than the Original Version this time……. Wish you would have added that removable battery

  • john adam belhami

    hoochie coochie

  • What is the best place to buy a gs3 for import to us?

  • Nowtimed

    Two things learned:

    1. Galaxy S3 is a BEAST

    2. Galaxy S2 is STILL a beast (gap between it and htc’s newest is marginal)

    • Fury

      #2 is ridiculous.

      Samsung is amazing.

  • JDWM

    I would be very interested to see HTC One X’ international performance now it has 4.0.3

      • JDWM

        Thank you so much for your reply. Do you have any thoughts on the battery life? I have an HTC Desire AND TWO batteries and I am concererned that I will not be able to open the One X. So I am thinking of moving to S3

        • Hi JDWM.

          Honestly, as much as I would love to answer your question, it’s far too early to make bold assertions as to what the battery life will be like.

          That being said – the battery on the S3 is removable, whereas the battery on the HTC One X is not. This alone is a deal-breaker (or maker) for some.

          It’s best that you wait until we get the device to our lab to do more thorough testing in 3 weeks, and we will let you know!

          Stay tuned and thanks for visiting!

          • JDWM

            Thank YOU for responding. Great advice.

          • Hey JDWM,

            You’re very welcome. It’s no easy feat staying on top of all this content and comments, let me tell you. Thanks for checking out the site. We’ve got some great stuff in store. Stay tuned!

  • Delroy_n_miller

    Why HTC One S wasn’t included in these benchmark tests? I saw a review that suggested that the One S performed better than the One X despite it’s lower specs.

    • Su_room

      Its cuz of the resolution

      • Justamazing87

        The galaxy s2 has lower resolution and was included. gl bench mark are not as reliable as quadrant or antutu…. quadrant had its issues before but not its very relliable.

        • Gameristhename

          I was saying for one s vs one x not s2 vs one s even if one s outperforms one x it will be because of resolution but s3 and.onex have same resolution and s3 outperforms , even the one with lower resolution.

  • ex.consumer

    I ran Vellamo on my Rezound with the Android 4.0.3 ‘leak’ and it came in at 1564, kind of better than this shootout’s 882! I’m starting to think these benchmarks are rather variable. In any case, this is all just entertainment to me, since the Rezound is the only phone I will blow my Verizon upgrade on for years to come, and its real-world performance is rather satisfying right now.

  • Rvbugged

    Samsung you still suck… Sorry!!

  • Ananyabhardwaj22

    Not worth the upgrade from galaxy s2

  • Susmacher

    If you have a galaxy s 2 stay with it the gap is not that big ,wait for the next galaxy this one is pretty impressive but it feels like a galaxy s 2.5 instead of a real breakthrough.

  • Mannieoosso

    Does it work as a phone that’s all I need to know

    • Kalliban

      Not out of the box, you will need to buy the phone-dongle for that!

  • Nojunkmail

    I love you Samsung Galaxy Slll

  • Duttaapd2009

    Not at all I expected better though I will surly buy this phone

  • Dmrulez_heartbreakkid

    Galaxy s3 rapes other phones :P

  • Geohound

    is this the skyrocket that the S3 is being compared to or the original S2 with the 1.2 ghz dualcore? I’m thinking I’m going to upgrade to the S2 skyrocket from my 3GS if that’s true since I understand I’m looking at benchmarks of a quadcore in the S3. Let me know if you think this is a bad idea and why.