Intel Core i9-14900K Tested in Geekbench & CPU-Z

[T0@st]

An alleged Intel Core i9-14900K engineering sample CPU was tested out recently in CPU-Z, with results leaked onto the internet earlier this week—courtesy of wnxod—978 points in single-core and 18117.5 points in multi-core. This particular sample of the flagship Raptor Lake Refresh processor managed to surpass its predecessors quite handily—with 9.7% SC/8.4% MC gains over the i9-13900K (Raptor Lake), and an uplift of 19.4% SC/59% MC over the i9-12900K (Alder Lake). Thanks to the i9-14900K's Thermal Velocity Boost (TVB) capability, it is able to hit a maximum 6.0 GHz clock speed (with P-cores) on 1.385 volts according to the leaked CPU-Z info.

Another example was put through the ringer via Geekbench 6.1.0 on Microsoft Windows 11 Pro 64-bit. The database entry popped up this morning, and several PC hardware news outlets were quick to pounce on the figures. In terms of single core performance, the benched Core i9-14900K achieved a score of 3121—blowing past a previous record holder—3089 set by a Core i9-13900KS CPU. Intel's 14th generation contender looks to be the fastest single-threaded chip out there, despite a less than optimal test system configuration—16 GB of DDR5-4800 memory on a Biostar Z790A-Silver mainboard, with Windows running a balanced power plan. The Core i9-14900K's multi-core score lagged behind its main rival—19032 versus 21678 (respectively). It would be nice to witness some nicer test builds materialize as we get closer to Intel's Innovation September event, and the rumored launch of K-series Raptor Lake Refresh processors around late October.




The Geekbench database entry and VideoCardz comparison chart are visible below:



View at TechPowerUp Main Site | Source

 

[Haile Selassie]

Assuming Package Power is closer to 400W than 288W, let alone '125W'?

 

[lolek86]

Even ryzen 7600 with safe auto +200mhz overclock and cheap 6000 memory with tight settings can do 3100 single core in geekbench. All that with only with single tower 4heatpipe Fera 5 cooling.

ASUS System Product Name - Geekbench

Benchmark results for an ASUS System Product Name with an AMD Ryzen 5 7600 processor.

browser.geekbench.com

 

[phanbuey]

Curious about the power as well. Although 15th gen looks to be right around the corner... not sure anyone on alderlake will want these unless the prices are low.

 

[P4-630]

not sure anyone on alderlake will want these unless the prices are low.

Well I won't be buying on release, that's for sure.

 

[_Under2World_]

Even ryzen 7600 with safe auto +200mhz overclock and cheap 6000 memory with tight settings can do 3100 single core in geekbench. All that with only with single tower 4heatpipe Fera 5 cooling.

ASUS System Product Name - Geekbench

Benchmark results for an ASUS System Product Name with an AMD Ryzen 5 7600 processor.

browser.geekbench.com

I don’t see any +200mhz but rather a base frequency of 7.73Ghz in the geekbench report. I don’t know if I get something wrong but it seems closer to +3Ghz vs base model.

 

[Onyx Turbine]

Even ryzen 7600 with safe auto +200mhz overclock and cheap 6000 memory with tight settings can do 3100 single core in geekbench. All that with only with single tower 4heatpipe Fera 5 cooling.

ASUS System Product Name - Geekbench

Benchmark results for an ASUS System Product Name with an AMD Ryzen 5 7600 processor.

browser.geekbench.com

so you mean a ryzen 7600 for much less money is almost as fast? i was under the assumption a ryzen 7600 is not even better as a 13500

 

[lexluthermiester]

Assuming Package Power is closer to 400W than 288W, let alone '125W'?

That would be a very poor assumption..

 

[Steevo]

That would be a very poor assumption..

Not really, the 13900KS pulls way over its rated TDP. Why would Intel give up their only lead by actually sticking to their TDP numbers?

 

[lolek86]

I don’t see any +200mhz but rather a base frequency of 7.73Ghz in the geekbench report. I don’t know if I get something wrong but it seems closer to +3Ghz vs base model.

Something craped out with geekbench under linux and this clock readings happend. It was precisely ryzen 7600 overclocked to 5.4GHz (It could get even higher, but is rather unstable without throwing stupid voltage and raising idle power consumption). Just look up 7600 in geekbench browser, you will find few users on tuned systems with 3100 single core

so you mean a ryzen 7600 for much less money is almost as fast? i was under the assumption a ryzen 7600 is not even better as a 13500

Ryzen 7000 likes tight memory timing. It really shows up in benchmarks compared to stock or even expo settings. Fortunately you can achieve this even with cheap 6000ddr5 sticks. I am using Veii M-die settings from ddr5 thread at overclockers forum. Combined with little overclock you can tune up to 15% higher single core compered to stock, propelling it almost to the top of single core charts.

 

[lexluthermiester]

Not really

Are you sure?

the 13900KS pulls way over its rated TDP.

Only when overclocked.

 

[R0H1T]

Not OCed, out of the box ~

 

[usiname]

Are you sure?

Only when overclocked.

Sure

 

[lexluthermiester]

Not OCed, out of the box ~

Sure
View attachment 312168

Fair enough.

 

[dyonoctis]

Not really, the 13900KS pulls way over its rated TDP. Why would Intel give up their only lead by actually sticking to their TDP numbers?

The 14900k is basically a 13900ks, so seeing it consume over 350w would be odd.

 

[Steevo]

 

[dyonoctis]

Sure
View attachment 312168

The i9 are funny because they got something called "TVB" were going over the PL2 (if the cooling allow it) is a feature. that's like zen4 X default behavior, without the benefits of the arch/node efficiency

 

[Crackong]

Having one component (GPU) drawing 400W+ in my system is troublesome enough.
There is no way putting another 400W monster in the loop.

I will stick to my 7800X3D.

 

[V3ctor]

When I see these numbers of consumption it reminds me of the PreHot times (Intel Prescot) Pentium 4, that all they did was increase Ghz and not the arch.

Hope we get another "Conroe" moment soon, because (I for sure) wont buy Intel products with such high consumption/thermals.

AMD right now is the way to go on efficiency

 

[ratirt]

It will consume more less the same as 13900 but will have some improvements. Maybe Intel will lower the consumption a tad and keep the performance the same but I'm pretty confident there would be not much benefit to change your current 13900K or KS to this one.

 

[JustBenching]

Having one component (GPU) drawing 400W+ in my system is troublesome enough.
There is no way putting another 400W monster in the loop.

I will stick to my 7800X3D.

Your 7800x 3d is slow. The 14900k will be miles faster than your 7800x 3d if you set both to same wattage, so I have no clue wtf are you trying to prove here

 

[Pinktulips7]

intel will destroy Ryzen or Dyzen with this, yes yes

Your 7800x 3d is slow. The 14900k will be miles faster than your 7800x 3d if you set both to same wattage, so I have no clue wtf are you trying to prove here

Yes Yes

 

[Denver]

Your 7800x 3d is slow. The 14900k will be miles faster than your 7800x 3d if you set both to same wattage, so I have no clue wtf are you trying to prove here

Generally, most here care about performance in games, where consumption is lower than in MT apps, but even so, X3D is more efficient.

 

[JustBenching]

Generally, most here care about performance in games, where consumption is lower than in MT apps, but even so, X3D is more efficient.

He mentioned 400w so I assumed he is talking about MT performance. Didn't realize he is trolling. Sorry

 

[lexluthermiester]

Your 7800x 3d is slow. The 14900k will be miles faster than your 7800x 3d if you set both to same wattage, so I have no clue wtf are you trying to prove here

What?!? What the hell are you talking about? The 7800X3D is THE gaming CPU to have. The only reason it isn't top-dog is that a few other CPU's do a bit better, but not by much and they are NOWHERE near the same level of value.

Yes Yes

Not really.

Didn't realize he is trolling.

There is a difference between trolling and expressing ones personal perspective. You might want to look into that.