MediaTek's new MT6592 octa-core chipset scores over 32,000 on AnTuTu
As our regulars would know, rushing into conclusions is something we try very hard to stray away from, at least until we get to review whatever needs reviewing first. It goes without saying that MediaTek's two new chips are rather exciting, as it would appear that both the successor to the quad MT6589T – the MT6588 – and the all new octa-core MT6592 have both crossed into hostile territory. This means that the performance on both these chips appears to be finally launching a credible threat to Qualcomm's complete dominion in the higher end. We gather that from the first few official benchmarks of the octa-core chipset, the higher-powered 2GHz version of which scores as high as 32,600. The less potent, 1.7GHz version of the MT6592 scores a tad lower, at just shy of 30,000 as you can see. Sure, syntethic benchmarks like AnTuTu are by no means and exhaustive performance criteria, yet they often provide a good perspective.
MediaTek did not disclose whether these scores resulted on a 720p or a 1080p device – an important distinction. Fortunately for us, the 1080p Zopo ZP998, a recent MT6592-touting entrant, has already been put through the test and scored 28,118.
Of course, there is some talk that AnTuTu is overly generous with its scoring on multi-core platforms, and we could definitely imagine that the software may not yet be fully optimized. Regardless, if you also consider the fact that the MT6592 comes with a quad-core Mali-450 GPU clocked at 700MHz, (@600MHz in the case of the new quad MT6588) you can definitely see how the new crop of MediaTek chipsets may be a threat to the status quo.
As always, we'll have to wait until a phone sporting the new chipsets graces our office before crowning MediaTek an official contender. Until then, the MT6592 only sounds great on paper.
1. _Bone_ (Posts: 2132; Member since: 29 Oct 2012)
Forget AnTuTu already for crying out loud. It's unreliable and benches a whole other things besides CPU, so it's not really useful since GPU, memory etc. can make a big difference.
4. Chris.P (Posts: 291; Member since: 27 Jun 2013)
Tests those as well. How well is a different question -- which is why we always cross-bench.
I'd actually love to hear your reasoning. For me, all this AnTuTu-bashing turned into a myth a while ago, as nothing of concrete substance was ever brought up. At least to my knowledge.
7. _Bone_ (Posts: 2132; Member since: 29 Oct 2012)
Results differ wildly on a single device, even more so on a different kernel, but that's nothing compared to the Nexus 5 doing half the AnTuTu scores of the "optimized" Note 3 on the same SoC... it's useless for proper CPU measurements, there are specific benchmark tools for that.
Of course it's a criticism of the manufacturers not of PhoneArena that's merely reporting the news. We just didn't learn a thing of the MediaTek performance because of the wrong benchmark tool was used.
13. Chris.P (Posts: 291; Member since: 27 Jun 2013)
It does differ wildly, indeed -- on our LG G2, we were getting scores with a margin of +/- up to 2,000 points at the extreme. This is much less pronounced on the lower/mid range of the spectrum, or when you go ahead and benchmark it a finite number of times, enough to strike a middle.
The point is, the scores fluctuate for all devices, so in a way it evens out at the end. Is it perfect? Heck no, but it does provide a now widely-recognized (!!!) platform by which to judge how well the device will handle itself given a set of granted circumstances. In fact, in the years I've used AnTuTu, I've never once had a score that was clearly inaccurate. Sure, It'll be impossible to call a 14,000 score device from, say, a 15,000 one, but you can still glean viable information from it. Which is why we still use it.
In other words, I still don't get what's so wrong with AnTuTu.
15. cezarepc (Posts: 570; Member since: 23 Nov 2012)
Nothing wrong with Antutu IMO too, but instead of using the result of a single device tested multiple times I find it more accurate to get the average scores of all the people who benchmarked their device (same unit).
I know a lot of people want to know the performance of a chip but as we've seen that hardly matters (some flagships are equipped with exactly the same internals but differs a lot in benchmarks). What matters is the end product where these chips ends up in.
Since every user will have different settings enabled, different amount of storage used, different amounts of remaining battery, temperature, etc that would give the overall world performance and a truelly more accurate representation of the device itself and not the chip inside it.
16. TediT (Posts: 37; Member since: 03 Sep 2012)
Please do explain why Nexus 5 is getting such low scores even though it has the best chipset on the market?
17. Taters (Posts: 3218; Member since: 28 Jan 2013)
Basically because pure android is not optimized for benchmarks, they don't care about it and possibly some thermal throttling going on.
The main reason though is because there are too many Ifans or HTC fans in the media and they perceive the Nexus 5 as too big of a threat. If the Nexus 5 crushed everything in benchmarks at $349 a pop that would be too powerful.
So all the ifans with media power concentrate on single thread scores and browser benchmarks masked as CPU benchmarks because they know chrome automatically loses to safari and they act like it is fair. Basically everything is abitrary and their is a lot of cherry picking to discredit the Nexus 5. Such as most reviewers giving the Nexus 5 bad scores for not having the best camera while the HTC One gets a free pass for having a bad one. Or all of a sudden the use of plastic over metal automatically deducts your score by 1 or 2. It is all arbitrary and Apple has too many converts in place.
18. Chris.P (Posts: 291; Member since: 27 Jun 2013)
As far as I remember from our Nexus 5 review, it wasn't cheating on the benchmarks like, say LG, Samsung or HTC. Read more here:
Moreover, thermal throttling ensures that the chipset never gets hotter than x degrees, which is apparently a bit too conservative on the N5 for an inefficient design or whatever reason.
19. picka_vi_materina (Posts: 164; Member since: 21 Nov 2011)
Since the phone sells for cheap, it must use much lower quality variants of the same S800 chip. Due to the lower quality part, heavy thermal throttling, just like the Nexus 4 is present in order to save the chip and provide longer lifetime.
11. silencer271 (Posts: 201; Member since: 05 Apr 2013)
They wont they love how their bosses iphone preforms on it
2. shahrooz (Posts: 165; Member since: 17 Sep 2013)
China is going to dominate Cellphone Industry with good quality yet not expensive phones in the near future. Mark my words!
14. Shatter (Posts: 2030; Member since: 29 May 2013)
Mediatek just needs to built US bands + LTE into their chips and it will happen.
3. mr.techdude (Posts: 551; Member since: 19 Nov 2012)
Most cellphones sold are Asian manufacturers, your prediction is more then halfway right.
5. cezarepc (Posts: 570; Member since: 23 Nov 2012)
If those benchmarks is an indication of how well these chips will perform it would be great to see phones and tabs rocking this. Lower price than Nexus devices, decent chips, well-known manufacturer. Hard to say no to that.
6. veer.d (Posts: 85; Member since: 12 Jul 2013)
I don't think it is threat to threat to qualcomm in higher end because it may not support LTE
9. itsdeepak4u2000 (Posts: 2689; Member since: 03 Nov 2012)
I think it still lacks some of the features which Qualcomm have.
10. taikucing (unregistered)
It even lacks compass and the GPS is crap. I don't even mention LTE
12. Chris.P (Posts: 291; Member since: 27 Jun 2013)
Has compass, GPS is hopefully fixed once and for all.
As far as I know, MTK is working on LTE already, though that's a considerable drawback, indeed. Then again, do consider that much of the world can still only wish for 4G.