Intel Core Ultra 7 270K Plus review: Back from the brink

Unbelievable productivity performance at an even more unbelievable price – shame gaming still takes a backseat.

Reviews
By published
Intel Core Ultra 7 270K Plus
(Image credit: © 3DTested)

Why you can trust 3DTested Our expert reviewers spend hours testing and comparing products and services so you can choose the best for you. Find out more about how we test.

Intel refocused on efficiency with Arrow Lake, and gave up performance in the process. Some of the efficiency advantages of the architecture are still in-tact, though Intel clearly pushed the power envelope to squeeze out as much performance as possible this time around. We’re still well below Raptor Lake levels of power demand, but Intel has pushed the Core Ultra 7 270K closer to that range.

Image 1 of 11
Intel Core Ultra 7 270K Plus
(Image credit: 3DTested)

Idle consumption is up, as well, with the Core Ultra 7 270K Plus consuming the most power out of our test pool in both true idle and active idle scenarios. The active idle situation is particularly surprising, as even Raptor Lake chips are able to maintain a lower power state than Zen 5 offerings in these low-lift situations.

Image 1 of 6
Intel Core Ultra 7 270K Plus
(Image credit: 3DTested)

The Core Ultra 7 270K Plus consumes a lot more power than similar CPUs around $300, but it also offers a lot more performance, at least in productivity apps. Looking at efficiency, Intel has sacrificed a bit compared to the Core Ultra 7 265K in Handbrake, though it’s still far more efficient than any Raptor Lake chip. AMD still rules the roost in efficiency, and the Core Ultra 7 270K Plus only furthers that narrative.

Image 1 of 3
Intel Core Ultra 7 270K Plus
(Image credit: 3DTested)

A better way to visualize efficiency is on a scatterplot, which you can see above. When reading these charts, the bottom right corner represents the best efficiency, while the top left corner represents the worst. These charts paint a clear picture of how Intel is pushing the Core Ultra 7 270K Plus. It draws a lot of power, but also delivers performance in equal measure in the applications we looked at.

Test Setup

All of our power measurements were gathered with a PMD2, which is a hardware monitor that captures power between the power supply and motherboard. We capture software metrics, as well, though only for comparison’s sake. The power measurements above are accurate, captured from the physical connectors themselves rather than the estimates provided via software.

For the test benches, they’re nearly identical, short of the motherboard and processor. The major difference is memory speed. For Intel platforms, we use 7200MT/s memory, as there’s a slight performance benefit with faster memory. AMD chips still run on 6000MT/s, as that’s the sweet spot for performance on Zen 4 and Zen 5 chips.

In addition to the same hardware, we use a frozen OS image with the same drivers and software to keep testing consistent. In the BIOS, we disable any automatic overclocking features (including AMD’s PBO as it isn’t warranted), as well as turn off VBS, turn on EXPO/XMP, and enable Resizeable BAR. We test on Windows 24H2, as 25H2 doesn’t offer any significant differences in performance.

Swipe to scroll horizontally

Intel LGA 1851 (Arrow Lake)

Row 0 - Cell 1

Motherboard

ASRock Z890 Taichi

RAM

2x16GB G.Skill Trident Z Neo RGB DDR5-7200

Intel LGA 1700 (Raptor Lake, Alder Lake)

Row 3 - Cell 1

Motherboard

MSI MPG Z790 Carbon Wi-Fi

RAM

2x16GB G.Skill Trident Z Neo RGB DDR5-7200

AMD AM5 (Zen 5, Zen 4)

Row 6 - Cell 1

Motherboard

MSI NPG X870E Carbon Wi-Fi

RAM

2x16GB G.Skill Trident Z Neo RGB DDR5-6000

All Systems

Row 9 - Cell 1

Gaming CPU

Nvidia GeForce RTX 5090 Founder’s Edition

Application GPU

Nvidia GeForce RTX 2080 Ti Founder’s Edition

Cooler

Corsair iCue Link H150i RGB

Storage

2TB Sabrent Rocket 4 Plus

PSU

MSI MPG A1000GS

Other

Arctic MX-4 TIM, Windows 11 Pro, Alamengda open test bench

TOPICS
Jake Roach
Jake Roach
Senior Analyst, CPUs
27 Comments Comment from the forums
  • Gururu
    I'd hate to be coming back into hardware scene right now. So many different SKUs from everyone, quite overwhelming.
    Reply
  • bit_user
    I think it's a little controversial to include iBOT in a hardware review, unless you at least test with it both on & off, so see how much it's contributing.

    I'm not really surprised to see something like this come along. I figured we'd have it by now, but I thought it'd be accompanied by hardware changes that required it. Based on my understanding, it's not really different than what JIT-based emulators are doing, for instance like when you run x86 code on ARM CPUs. In this case, it just so happens to be doing x86 -> x86.
    Reply
  • TerryLaze
    Admin said:
    Gaming performance still struggles,
    114 minimum FPS on average in a suite of 17 games...."Struggles"
    Reply
  • colossusrage
    TerryLaze said:
    114 minimum FPS on average in a suite of 17 games...."Struggles"
    Yeah, poor choice of words, maybe struggles to keep up with 9800X3D, but on its own it's a good gaming CPU.
    Reply
  • usertests
    The die-to-die frequency increase has helped it to perform much better than a typical refresh, although it clearly tanks efficiency and idle power consumption badly.

    Combined with the price, while it's not magic, it's the best possible outcome for Arrow Lake.

    It will be interesting to see if the other reviews are so generous with iBOT.
    Reply
  • Gururu
    Reviews across the board are painting it as an absolute best for value. HU complained about the temps but the major tiff everyone had was of course the platform, being DDR5 and dead end. If I waited this long to upgrade from a DDR4, I'd just wait for Nova or Zen 6. If I have an 1851 already, the performance bump doesn't warrant more spending. If I was buying for family or significant others who don't upgrade period, this is a no-brainer.
    Reply
  • Notton
    I know builder and tycoon games aren't popular, but if you really want to test out CPU performance, load an end game save from Factorio, Timberborn, Cities Skylines 2, or Transport Fever 2.

    The path finding calculation will bring a 9850X3D to its knees, and you'll get to see the true value of an X3D processor.

    Also, where is the i5 250K review?
    Reply
  • warezme
    TerryLaze said:
    114 minimum FPS on average in a suite of 17 games...."Struggles"
    This is valid. The test configuration is a 5090FE at 1920x1080, where the CPU becomes the limiting factor. Otherwise if you were to put a 5090FE on pretty much any modern CPU you would get equally high frame rates even at higher resolutions just not as many as you would on a 9800X3D, 9850X3D or even 9700 where the 270 wouldn't keep up.
    Reply
  • rluker5
    warezme said:
    This is valid. The test configuration is a 5090FE at 1920x1080, where the CPU becomes the limiting factor. Otherwise if you were to put a 5090FE on pretty much any modern CPU you would get equally high frame rates even at higher resolutions just not as many as you would on a 9800X3D, 9850X3D or even 9700 where the 270 wouldn't keep up.
    No Ryzen that isn't an X3D tested here can keep up with the 270K+ in games, the more expensive and vastly slower in everything else 9700X included.
    Https://cdn.mos.cms.futurecdn.net/WVvc7x7yHrYgJKwdTp7WEo-1200-80.png.webpTom's should just add that the non X3D Ryzens are all a worse choice for gaming when they mention that better gaming chips are lackluster. That would be a way to seem unbiased.

    Edit: You could call the 270K+ the $300 9950X.
    Reply
  • bit_user
    Gururu said:
    If I waited this long to upgrade from a DDR4, I'd just wait for Nova or Zen 6.
    But, what if they're both delayed until 2027, as some rumors have suggested?
    Reply