Which CPU should you buy?? 14600K vs 13600K vs 12600K

Intel’s 14th gen CPUs are here and… well… they’re pretty much just the 13th gen chips with a new name. Seriously. It’s the same architecture, it’s Raptor Lake-S, it’s the same process node, Intel 7, and with this 14600K in particular there is very little different. It still has 14 total cores, 6 P and 8 E cores, it still has 24MB of total cache, it still boosts to a maximum of 5.3GHz, it still has a max turbo power of 181 watts. The only spec sheet difference I can find is that the E cores can now boost to 4GHz instead of 3.9GHz. GAME CHANGING, I KNOW! There really isn’t much new – there’s a few bits like a new “AI assistant” built into their extreme tuning utility software, and “Application Optimisation” or APO which basically does some under-the-hood tweaks based on the game’s executable name, but realistically both of those things could work on older chips, but Intel obviously wants to keep them for the new stuff to make it worth you buying the newer chip.

Of course, there is always more beyond just the specs, so let’s dive straight into the benchmarks, and this time I’m gonna start with the gaming results. I should note here that for a number of reasons, including it being a realistic choice for these chips and because I don’t have anything better, I’m using an RTX 3070 Ti for this testing. I’m testing at what I’d consider “realistic” settings too, and using the best kit of RAM I have access to, that being Kingston Fury DDR5-5600 CL40, and I’m using the latest public BIOS version of the Asus Z790 Strix E and otherwise stock BIOS settings, save for using the XMP 1 profile and Dynamic Tuning Technology enabled. That does mean the power limits are removed, but that’s the stock setting on this board and is how I’d imagine most people will be using these chips. It’s the same board, RAM, cooler and GPU for all these chips, so hopefully that’s controlled enough for ya. Right, games!

I’ll start with the new kid on the block, CS2. Now it’s worth pointing out that CS2 is still very much in the “public beta” stage, as much as Valve would have you believe otherwise, so take these results with a pinch of salt. With that said, while the 13th and 14th gen chips do hold a comfortable lead over the 12th gen i5, the gap between 13th and 14th isn’t nearly as big, although it is there. All of these results are more than playable though. At 1440p the gap to the 12600K remains roughly the same, but the distance between the 13600K and 14600K shrinks to under 10 FPS – and at over 400 FPS average, that’s not what you’d call “noticeable”.

Cyberpunk on medium settings at 1080p is the same story. The 12600K is about 25 FPS slower than the 13th and 14th gen chips, while those two only share a 2 FPS difference. There is no significant improvement between the now previous generation and this “new” one. As expected, 1440p flattens those results even more, with only a 3 FPS spread across all three chips. There is a notable difference in the 1% low figures though, with the 12600K running 10 FPS lower than the other two, meaning you’d get a smoother experience on the 13th and 14th gen chips.

Microsoft Flight Simulator is a really interesting one, as the 14600K actually ran 6 FPS slower than the 13600K. I made sure to retest this but at least with my setup it was a repeatable result. It’s hardly a big difference, and Intel’s own slides show some games have a bit of performance regression so I’m not too worried. Of course the 12600K is at the back of the pack, running 15 FPS slower than the 14600K and 20 FPS slower than the 13600K. At 1440p things are back to normal with the 14600K in the lead – just – and the 12600K running 15 FPS behind. 

Fortnite at 1080p on the high preset with no TSR was pretty strange – basically all the results were capped at 106 FPS, give or take, err, half a frame? Yeah, pretty weird. 1440p isn’t much better, again all within 1 FPS of each other, although strangely the 13600K has slightly lower 1% lows in both 1080p and 1440p results. 

Hitman 3 is great because it breaks out the CPU and GPU performance, so this is the CPU results, and at medium settings at 1080p we can finally see a noticeable improvement from the newest 14th gen chip. It isn’t exactly much, it’s under 10 FPS, but it’s an improvement nevertheless. The gap to the 12600K is still there too, with a little over 15 FPS between that and the 13600K. 1440p brings all that down though, where the 14600K and 13600K are really very close – just 3 FPS between them. Even the 12600K has caught up, now only 10 FPS or so behind. 

Rainbow Six Siege is back to normal, with the 14600K just ahead of the 13600K, followed by a fairly large gap to the 12600K. All of these results are more than playable though, so I wouldn’t exactly be worried if you have a 12600K. At 1440p that gap is even closer, with the 12600K just 20 FPS behind the 13th gen chip, and just a 5 FPS gap to the 14600K. Really not much in it. 

Shadow of the Tomb Raider is the same story. At 1080p there’s only 2 FPS on average between the two Raptor Lake chips, or 4 FPS in the 1% lows. The 12600K does trail by the usual margin, running around 20 FPS slower, but at 180 and 200 FPS respectively, I can’t say that’s a “deal breaking” amount of difference. At 1440p they may as well all be the same chip. Seriously. The 1% lows AND averages are functionally identical, showing a complete GPU bottleneck. 

Lastly, in Starfield on the low preset – because how else can you play this damn game – there’s a reasonable amount of differentiation. The 13600K is lacking in the 1% lows again, although on average it’s pretty close to the newer 14600K. The 12600K trails the usual amount, this time only around 10 FPS average compared to the 13600K, and with actually better 1% lows than the 13600K. At 1440p it’s a similar story, albeit with less performance on tap. The 1% lows on the 13600K are still a little low compared to the other two chips, but otherwise it’s pretty close. 

To summarise all those results, here’s the percentage differences at 1080p comparing the 14600K to the 12600K. On average, the 14600K is about 17% faster, with the highlights being CS2, Hitman 3 and Rainbow Six Siege. Now if I swap that to compare the 14600K to the 13600K, that’s a lot less impressive. On average the 14600K is just 1.9% faster across these 8 games. That’s… abysmal. At 1440p the 14600K is 9.4% faster than the 12600K, and just 1.1% faster than the 13600K. Yikes.

Of course we need to talk about the CPU workloads, so let’s take a look at those, starting with Cinebench. Single threaded performance has improved ever-so-slightly, which adds up to under 1000 points more in all-core work. It’s nowhere near the improvement the 12th to 13th gen saw, as you should probably expect by now. Blender is the same, a frankly marginal improvement from 13 to 14, but a significant, generational improvement from 12 to 13. Puget bench for the Adobe apps – namely Premiere Pro, After Effects and Photoshop all show the 13th and 14th gen chips functionally identical, while the 12600K trails behind, somewhat significantly in fact. Lastly, looking at power consumption, both the stable on-load power, and the peak power draw, the highest I saw was 155W on the 14600K, or 153W on the 13600K. Really, a slight overclock to the 13600K would provide all the performance (and then some) that the 14600K provides. In CPU specific tasks, the 14600K is just 2.6% faster than the 13600K. 

So, what the hell is going on here? Well, if I’m being blunt, this is a shareholder launch. This isn’t a “look we made this awesome new thing you should buy” launch, this is a “we are a massive multi-billion dollar corporation and if we don’t keep to our yearly launch cycle our stock price is going to drop and our execs won’t get their stock price related bonuses” launch. There is functionally nothing new here, save for the i7 which now gets more E cores. That’s about it. This is a stopgap until they can get the legitimately interesting stuff – Meteor Lake – into production. So, what chip should you buy? I guess either the 13600K or 14600K, whichever is cheaper, because they’re the same thing. These new chips don’t even require a BIOS update to work. That tells you all you need to know.