The Best CPUs for 2025

Deeper Dive: Our Top Tested Picks

Pros & Cons

• High core count for the money
• Competitive multi-threaded performance

• No integrated graphics in “F” version
• Lackluster single-threaded performance
• Somewhat loud stock fan

Why We Picked It

Driving better performance per dollar than most other Intel CPUs, the Core i5-13400F is a standout option in Intel’s desktop lineup. It lacks integrated graphics, but Intel makes up for this by dropping the price. If you plan on buying a graphics card, though, this is arguably well worth the trade and can help you save a few bucks on the processor that could be put to use instead toward more RAM or a faster video card. (If you need the integrated graphics, opt for the non-“F” Core i5-13400 for a little more money.)

Who It’s For

For anyone building a midrange gaming PC, the Core i5-13400F should be one of the top options you consider. Faster processors are always available, but you aren’t likely to need more speed in a midrange gaming PC than what the Core i5-13400F can muster. Look for something else, though, if you aren’t already planning to buy a graphics card, as the no-graphics “F” savings aren’t enough to justify buying a graphics card for that reason alone. Yes, a higher-end processor than this may better extend the useful life of your PC, but the Intel Core i5-13400F is hard to beat for its price.

Specs & Configurations

Core Count

10

Thread Count

16

Base Clock Frequency

2.5 GHz

Maximum Boost Clock

4.6 GHz

Socket Compatibility

Intel LGA 1700

Lithography

10 nm

L3 Cache Amount

20 MB

Thermal Design Power (TDP) Rating

148 watts

Integrated Graphics

None

Integrated Graphics Base Clock

Bundled Cooler

Intel Laminar RM1

Learn More

Intel Core i5-13400F Review

Pros & Cons

• Strong performance
• Better value than Ryzen 7 5800X
• Compatible with most AM4 motherboards
• 65-watt TDP

• Falls behind the more affordable Core i5-12600K

Why We Picked It

Yes, technically, the AM4 platform is old hat now, but it still has life in it for budget shoppers. The Ryzen 7 5700X was introduced toward the end of the AM4 platform’s life as AMD’s flagship mainstream processor, and it launched with a price that undercut many existing Ryzen 7 5000-series chips. Despite its lower price, this eight-core CPU drives similar speeds to most other Ryzen 7 5000-series processors. The lower price without any appreciable change in performance made it an alluring option at launch, and this still holds today, especially as its price has dropped further.

Who It’s For

If you own an AM4 motherboard already, this is a great budget chip option. As mentioned above, the Ryzen 7 5700X can be picked up today for less than its entry price, but even back at its launch time it was already a highly attractive option. Its performance is among the best of the Ryzen 7 5000 series, and it manages games in a particularly excellent fashion for the money. This makes it one of the best options to consider if you want to build a no-fuss gaming PC and save a few bucks on the processor, which will help you stretch your budget for a better graphics card.

Specs & Configurations

Core Count

8

Thread Count

16

Base Clock Frequency

3.4 GHz

Maximum Boost Clock

3.6 GHz

Socket Compatibility

AMD AM4

Lithography

7 nm

L3 Cache Amount

32 MB

Thermal Design Power (TDP) Rating

65 watts

Integrated Graphics

None

Integrated Graphics Base Clock

Bundled Cooler

None

Learn More

AMD Ryzen 7 5700X Review

Pros & Cons

• CPU performance punches above its station in bench testing
• Reasonable price-per-core
• High overclock ceiling

• Testing games with integrated graphics was bumpy, and slower than AMD’s competition
• Z690 platform demands high cost of adoption, versus CPU purchase price
• At launch, not compatible with some games that use Denuvo DRM protection

Why We Picked It

Yes, this processor is a few generations old, but that doesn’t mean the Core i5-12600K still can’t perform adequately in many scenarios. This chip has six Hyper-Threaded CPU cores that boost up to 4.9GHz, more than sufficient power for office work or everyday tasks, like heavy multitasking and web browsing. As it’s grown older, this CPU’s price has dropped, and so has the cost of motherboards that support it, which makes it a more affordable option than many newer solutions.

Who It’s For

Though no longer a top option, the Core i5-12600K was popular for gaming PCs at its launch, and it can still serve adequately in this capacity for a budget gaming PC. It is likewise a worthwhile option for businesses and families to consider. Realistically, buying a newer solution would probably be best for most folks. Still, the Intel Core i5-12600K is a compelling value if budget primarily drives your hardware purchases. (Note that the Core i5-12600K has a workable Intel UHD Graphics built-in display solution. If you don’t have a graphics card, avoid the similar-sounding, slightly cheaper Core i5-12600KF, which has the on-chip graphics disabled.)

Specs & Configurations

Core Count

10

Thread Count

16

Base Clock Frequency

3.7 GHz

Maximum Boost Clock

4.9 GHz

Socket Compatibility

Intel LGA 1700

Lithography

10 nm

L3 Cache Amount

12 MB

Thermal Design Power (TDP) Rating

125 watts

Integrated Graphics

Intel UHD Graphics 770

Integrated Graphics Base Clock

300 MHz

Bundled Cooler

None

Learn More

Intel Core i5-12600K Review

Pros & Cons

• 16-core design
• Runs significantly cooler than its 7950X predecessor
• Marked performance gains
• Cheaper than last-generation flagship

• Power-hungry
• At launch, a little pricier than the Core i9-14900K

Why We Picked It

If you want to build a PC around an AMD processor capable of the best possible performance that AMD has to offer, regardless of the cost, then the Ryzen 9 9950X is the right option for you. At the moment, it is unquestionably AMD’s fastest consumer-oriented processor, packing 16 CPU cores, 32 threads, and a blistering-fast 5.7GHz clock speed. Based on the Zen 5 microarchitecture, the Ryzen 9 9950X succeeds in completing more work each clock cycle than its predecessor, giving it better performance in most tasks even while running at the same clock speed.

Who It’s For

The Ryzen 9 9950X is the best option from AMD for you if you do a lot of work that is CPU-intensive. This includes compiling software, heavy multitasking, and content creation work, like rendering videos and editing images. The Ryzen 9 9950X could be an option for a high-end gaming PC, too, but realistically, it’s overkill for that purpose in most scenarios, unless you use the same PC for professional content work, too. And if that’s the case, you may want to consider the Ryzen 9 9950X3D instead.

Specs & Configurations

Core Count

16

Thread Count

32

Base Clock Frequency

4.3 GHz

Maximum Boost Clock

5.7 GHz

Socket Compatibility

AMD AM5

Lithography

4 nm

L3 Cache Amount

64 MB

Thermal Design Power (TDP) Rating

170 watts

Integrated Graphics

AMD Radeon Graphics

Integrated Graphics Base Clock

2200 MHz

Bundled Cooler

None

Learn More

AMD Ryzen 9 9950X Review

Pros & Cons

• Major performance increase in some tests
• Reduced power consumption
• Improved integrated graphics
• Dedicated AI silicon
• Worthwhile platform enhancements

• CPU-centric tests are a mix of wins and losses
• Subpar gaming performance in some titles

Why We Picked It

One of Intel’s fastest processors to date, the Core Ultra 9 285K has eight high-performance P-cores and 16 efficient E-cores with boosted performance over the previous generation. Intel dropped Hyper-Threading support for the Ultra 9 285K, but its large core count and architectural improvements enable this flagship chip to beat out the old Core i9-14900K even though it is clocked lower with a max turbo speed of 5.7GHz. The Ultra 9 285K is also less power-hungry than its predecessor, and it runs cooler, improving several areas beyond just performance. Finally, this chip has a reasonable price relative to its performance.

Who It’s For

Anyone who wants the best performance possible from a desktop PC for non-gaming tasks should consider buying an Intel Core Ultra 9 285K. While this processor is certainly capable of gaming, too, it falls behind its AMD competition in this area. With all other tasks, however, the Ultra 9 285K performs exceptionally well, showing top-tier content creation performance driven by its high core count.

Specs & Configurations

Core Count

24

Thread Count

24

Base Clock Frequency

3.7 GHz

Maximum Boost Clock

5.7 GHz

Socket Compatibility

Intel LGA1851

Lithography

3 nm

L3 Cache Amount

36 MB

Thermal Design Power (TDP) Rating

250 watts

Integrated Graphics

Intel Xe LPG

Integrated Graphics Base Clock

2000 MHz

Bundled Cooler

None

Learn More

Intel Core Ultra 9 285K Review

Pros & Cons

• Potent CPU performance
• Impressive gaming performance
• Reasonably priced
• 128MB L3 cache

• Biggest gains are in niche scenarios
• Not much faster than 9950X in CPU workloads

Why We Picked It

The Ryzen 9 9950X3D is AMD’s top consumer-oriented processor with the best performance it has to deliver in the category. Its 16 high-performance CPU cores support SMT technology and run at up to 5.7GHz. This speed enables it to run applications exceptionally fast, with few able to match it. This chip also uses AMD’s second-generation 3D V-Cache technology (thus the “X3D”), which boosts it for PC gaming for not much more cash than the standard Ryzen 9 9950X. The only aspect of this chip that isn’t cutting-edge is its integrated graphics portion, which can serve in a pinch if you lack a graphics card. (Given what this chip is for, though, most buyers aren’t likely to use the integrated graphics at all.)

Who Its For

To get the very most out of the Ryzen 9 9950X3D, you need to own one of the fastest graphics cards that money can buy, like an Nvidia GeForce RTX 5090. You can still benefit from the Ryzen 9 9950X3D with lesser graphics cards, but that depends on how fast your graphics card is and the kind of games you play. If you want to ensure you have the best that money can buy for a gaming CPU, then this is what you want.

Specs & Configurations

Core Count

16

Thread Count

32

Base Clock Frequency

4.3 GHz

Maximum Boost Clock

5.7 GHz

Socket Compatibility

AMD AM5

Lithography

4 nm

L3 Cache Amount

128 MB

Thermal Design Power (TDP) Rating

170 watts

Integrated Graphics

AMD Radeon Graphics

Integrated Graphics Base Clock

2200 MHz

Bundled Cooler

None

Learn More

AMD Ryzen 9 9950X3D Review

Pros & Cons

• Excellent performance
• Four more E-cores than predecessor
• Increased cache
• No price increase

• Power hungry
• Overheats with 240mm water cooler

Why We Picked It

Intel’s Core i7-14700K is one of the best overall processors that money can buy, with excellent performance in all areas. The i7-14700K delivers notable improvements over the Core i7-13700K, with four additional E-cores, raising the total core count to 20. The processor also has slightly elevated clock speeds compared with its predecessor, giving it a slight edge in single-threaded tests. The price isn’t half-bad either, undercutting the more costly Intel Core i9 options and AMD’s Ryzen 9 competition while providing almost the same level of performance as these higher-end chips in most situations. This processor’s gaming performance is also exceptional, rivaling the Intel Core i9-13900K in many titles.

Who It’s For

For pure gaming purposes, it can be a little difficult to justify the Core i7-14700K over the Core i5-12600K or the Core i5-13600K, which are lower-priced and produce comparable performance in games. In non-gaming tasks, however, the Core i7-14700K delivers a sizable performance advantage, making it a clearly superior choice if you plan to do more than just game on your PC. This can come in especially handy if you do things like record your gameplay, stream while gaming, or anything with that content in terms of rendering or encoding.

Specs & Configurations

Core Count

20

Thread Count

28

Base Clock Frequency

3.4 GHz

Maximum Boost Clock

5.5 GHz

Socket Compatibility

Intel LGA 1700

Lithography

7 nm

L3 Cache Amount

33 MB

Thermal Design Power (TDP) Rating

254 watts

Integrated Graphics

Intel UHD Graphics 770

Integrated Graphics Base Clock

1600 MHz

Bundled Cooler

None

Learn More

Intel Core i7-14700K Review

Pros & Cons

• Excellent multi-threaded performance
• 32 cores with 64 threads
• Faster clock speeds than Threadripper 7980X
• Solid price-for-performance equation

• Runs hot
• Needs registered RAM

Why We Picked It

AMD’s return to the high-end desktop (HEDT) market with its Ryzen Threadripper processors means that Team Red now technically offers the fastest single processors for consumers. (Its Threadripper Pro chips of the last two generations were primarily enterprise fare.) These chips’ huge core counts often make them multiple times faster than the closest competing processors on other consumer platforms for certain heavily multithreaded tasks. The Ryzen Threadripper 7970X, in particular, stands out on value. At $2,499, it’s half the price of the Ryzen Threadripper 7980X while offering still-mighty performance. It doesn’t have quite as many cores, which hurts it in heavily threaded applications versus the 7980X, but its higher clock speeds partially make up for this deficit in multi-threaded workloads, and give it an advantage when fewer than 32 cores are in action. That’s going to be most of the time, unless you are running heavily threaded applications like software compilers or image-rendering programs. Even then, unless you do that a lot, the Threadripper 7970X’s price advantage may make it a better choice.

Who It’s For

Make no mistake: These aren’t casual-user or gamer chips. The 7970X is best suited for extreme multitasking on multiple demanding applications. Like all Threadrippers, its true potential comes out in heavily threaded programs like Photoshop and other content-creation fare, as well as for CAD work, scientific analysis and modeling, or advanced research and development applications.

Specs & Configurations

Core Count

32

Thread Count

64

Base Clock Frequency

4 GHz

Maximum Boost Clock

5.3 GHz

Socket Compatibility

AMD sTR5

Lithography

5 nm

L3 Cache Amount

128 MB

Thermal Design Power (TDP) Rating

350 watts

Integrated Graphics

None

Integrated Graphics Base Clock

Bundled Cooler

None

Learn More

AMD Ryzen Threadripper 7970X Review

Get Our Best Stories!

All the Latest Tech, Tested by Our Experts

Sign up for the Lab Report newsletter to receive PCMag’s latest product reviews, buying advice, and insights.

By clicking Sign Me Up, you confirm you are 16+ and agree to our Terms of Use and Privacy Policy.

Thanks for signing up!

Your subscription has been confirmed. Keep an eye on your inbox!

Buying Guide: The Best CPUs for 2025

CPU Consideration No. 1: Goals and Upgrades

It’s important to set a goal for what you want to achieve with a change of desktop CPUs. You want better performance, but are you looking for gaming performance, specifically? Is it for a system you will use for day-to-day work? Or do you simply have an older computer sitting around that you’d like to spruce up for internet use and watching videos?

We can’t stress enough the importance of knowing what you want out of your system, to make sure you don’t end up with a disappointing amount of horsepower in the end, whether that’s due to overspending or underspending. This will also largely help you to determine if you want to upgrade your PC or build a new one.

(Credit: Michael Justin Allen Sexton)

First, we should be clear that upgrading is often a viable option, especially if you just want the system for light-duty use like web browsing. Technically, even antique PCs based on AMD’s Phenom, Phenom II, and FX series are still suitable for this task, as are Intel’s Core 2 Duo and Core 2 Quad systems, but these relics would certainly show a good deal of slowdown if you pushed them very hard. These aforementioned chips certainly aren’t worth upgrading or purchasing now—far better options are available for the money.

If your computer is relatively modern, though, it might be better for you to upgrade your current PC instead of buying a new one. This makes your choice of processor far simpler, as there will be a limited number of chips that will work with your current system. Upgrading isn’t always an option, as you might already have the best CPU that your current motherboard supports. But as long as that’s not the case, upgrading can be the most affordable way to end up with a faster PC, as you won’t need to purchase other components.

As to whether it’s worthwhile to upgrade your system, a good rule of thumb is to consider the age of the system you’re using. If that system was made within the last 10 years and doesn’t already have the fastest CPU available for it (or close to it), upgrading may get you the performance boost you desire. Nothing older than this is remotely worth upgrading, as it becomes more practical to save for a newer system instead. And even with a system that’s around 10 years old, it’s only worth considering if you have something like an old Intel Core i3 or worse, and you are upgrading to an old Core i7—but this is also only worth considering if you can find that latter processor for cheap.

A quick search shows that for older PCs based on Intel’s LGA 1155 socket, which was the platform used by Intel’s second- and third-gen Core “Sandy Bridge” and “Ivy Bridge” processors (now more than a decade old), you can buy second-hand Core i7-3770 processors for around $30 to $50 from take-your-chances sources like eBay. This processor today would lag behind a modern Core i3, but it would still be capable of web browsing or even use in a low-end gaming PC. Again, we must stress, however, that this sort of upgrade only makes sense if you have one of the slowest CPUs available for that old platform and are upgrading to one of the fastest ones. If you had a Sandy Bridge Core i5, it would make more sense to save for a new system instead of upgrading, since the ceiling for the best chip available on that platform isn’t much higher.

Upgrading more recent systems also follows this same logic, i.e. it only makes sense to upgrade if you are making a substantial jump on the same platform—and you can do so cheaply. Even on an almost new system, we would hesitate to upgrade from, say, an AMD Ryzen 5 9600X to an AMD Ryzen 7 9700X, or from an Intel Core Ultra 5 245K to an Intel Core Ultra 9 285K. It’s not that you would not see a noticeable and measurable performance boost (our reviews show that there clearly would be one), but the gains you receive for the cost of upgrading often don’t make it worthwhile. This makes it exceedingly important that you buy the right processor from the start, rather than waiting and trying to upgrade later.

If your goal is to simply fix up your old PC for web browsing or low-end gaming, then you may want to stop reading here and see if an affordable upgrade option is available for your system. Each motherboard, including the one in your PC, has a list of supported processors that you should be able to find on the manufacturer’s product page. If none of the processors supported on your current motherboard is affordably priced or a significant step up, then just forget about upgrading and start thinking about building or buying a new system instead. If your goal was instead to get a faster system for work or high-end gaming, or if upgrading just won’t cut it for you, then keep reading.

System Planning 101: How to Pick a Motherboard, a CPU, and RAM That Work Together

If you’re going to build a new system, then you’ll need to pick mostly new parts. If you had an older system, you might be able to re-use a few items, notably the power supply, case, or storage device, but everything else you’ll likely have to buy. Some parts, like the ones we just listed that you could reuse, are compatible with a wide range of systems and can last for several generations. But motherboards, CPUs, and main system memory (RAM) are the opposite.

(Credit: Joseph Maldonado)

These items will only last for a few generations, at most. All motherboards have a socket that the CPU is designed to be mounted in, and at the chip maker’s level, that socket’s design has to be changed every so often to allow for the addition of new features. As a result, motherboards only support CPUs that will fit in (and are explicitly compatible with) their sockets, and the same is true vice versa for CPUs. (Sometimes a chip will physically fit, but it isn’t supported.) RAM, similarly, has a custom set of slots that it mounts on a PC motherboard, and these slots only support one broad type of RAM and won’t work with any others. (The latest types are known as DDR4 and DDR5.)

(Credit: Joseph Maldonado)

Due to sockets and, thus, motherboard platforms changing every few years, you will typically need to buy a new motherboard and, possibly, a new set of RAM if you want to buy a new CPU that succeeds your current one by more than a few years. The only time this won’t be the case is if you are upgrading on the same platform, like in the possible scenarios discussed above. If you upgrade frequently, your old RAM might remain compatible with your new system.

In that vein, when buying parts, you’ll want to make sure the parts you get will be compatible with each other. CPU sockets are typically numbered; for mainstream CPUs bought by most consumers, the latest are AMD’s AM4 and AM5, and Intel’s LGA 1700 and LGA 1851, and both motherboards and processors will be identified by this number to make picking a matching pair easy. Still, you’ll want to make certain that a given CPU is explicitly supported by a given board; verify that compatibility on the board maker’s website.

(Credit: Joseph Maldonado)

For several years now, RAM has also followed a simple numbering scheme as we’ve progressed from DDR to DDR2, DDR3, DDR4, and now DDR5. None of these will work in RAM slots made for the other, and this is also detailed on the motherboard specs page, which again makes finding a matching compatible part relatively easy.

Power, Now! Don’t Build With Plans of Upgrading

Often, people consider building a PC with plans to upgrade it in the future. The idea is that if you get a motherboard and CPU now and can later upgrade to a faster CPU, this might help you avoid having to build another computer and save costs down the road the next time you feel your PC slowing down.

The problem with this train of thought? It rarely works out as planned. Intel, over the last decade, has changed to a new CPU socket every two or three years, typically once every two CPU generations. AMD postpones changing sockets longer, and it only recently introduced its AM5-socketed motherboards as a successor to its AM4 motherboards. AM4 was first launched in 2017, giving it a five-year reign during which time AMD pushed out four primary generations of processors for the platform. We don’t know how long AM5 will be AMD’s main platform at this time, but there’s good reason to believe it will be used for roughly the same period as AM4 was.

(Credit: Joseph Maldonado)

Even with AMD’s extended platform lifespan, you aren’t overly likely to upgrade processors on the same motherboard. Compatibility remains a question mark even with AM4, as not all processors made for the platform are supported on all motherboards. AMD made efforts, especially toward the end of AM4’s reign of dominance, to encourage board makers to push out updated BIOS versions for older motherboards to extend support for their newest AM4 CPUs to older motherboards. But this effort still relied on AMD’s board partners to implement the new BIOS versions on dozens or even hundreds of boards.

Long story short: You cannot buy a motherboard today and be absolutely confident it will support processors released a few years down the road, even if they are released for the same socket as the motherboard you are buying today.

Even if you can upgrade the CPU on an existing board, strong reasons may arise why you might not want to. Newer motherboards with updated chipsets have their inherent benefits. These might include things like support for faster RAM, faster USB and storage connections, faster PCI Express slots and underlying buses, faster internet support, and myriad other potential benefits. Plus, after you upgrade, you’ll be left with an orphaned CPU to then resell or put into a new motherboard. It starts to make a lot more sense to just think about selling (or giving away) your old PC or CPU/motherboard combo, and building new, most of the time, when you need something faster.

If you are buying into an Intel platform, this is doubly true. Unless you buy a low-end CPU, to begin with, you aren’t likely to want to upgrade in just one or two years before Intel introduces a whole new platform. Though there is some merit in the idea of doing just that (buying a low-end CPU to upgrade next year when high-end prices drop), it still raises overall costs, as you now have a low-end CPU you paid for in addition to a high-end CPU. You’d be better off just buying that high-end CPU to begin with and enjoy it that much longer.

All in all, though some may view platforms that are aging out or about to be replaced as dead ends, it’s best to view all platforms in this same frame of mind. That’s not to say that none of them is worth buying (that’s not true), but upgradability shouldn’t be first on your mind when selecting a motherboard and CPU to use in your new PC. Some things can be easily upgraded later, like the RAM, storage, or graphics card, but this is not an advantage that motherboards or processors have.

When you do buy a system, unless you are getting an exceptional deal, you should still buy into the newest platform available with the best processor you can afford that fits your needs. Currently, Intel’s newest platform is the LGA 1700 platform, and AMD’s (as mentioned above) is AM5. There’s no harm in getting an AM4 system or a last-gen Intel LGA 1200 system, but, unless you’re getting those steeply discounted from their launch prices, you’re going to get a better system and more bang for your buck by going with what’s newest.

Buying Basics: Four Key Concepts to Know About CPUs

Now that we’ve covered the basic considerations of whether to upgrade or buy new and the other hardware you’ll need to consider when buying a CPU, let’s talk about what differentiates one processor from another. A multitude of factors need to be considered here, but the most important ones are the microarchitecture, core count, thread count, and clock speed.

Microarchitecture

This is by far the most important aspect of any CPU, and indeed the same is true for just about any piece of technology. It’s the design and internal blueprints of how the device is built and what makes it work.

Due to differences in architecture, you can have multiple CPUs that operate at the same speed but still perform drastically differently. It might be helpful for you to picture each architecture in a similar vein as you might consider different makes of cars or airplanes. In the same way that you can have multiple cars with their engines running at 2,000rpm and get drastically different performance and speeds out of the cars, the same is true for CPUs.

(Credit: Joseph Maldonado)

Judging architectures is exceedingly difficult, as they are incredibly complex, with billions of transistors and a multitude of other external factors that bear on their performance. To gain an idea of how different architectures perform, you should read our reviews that touch on the subject and also compare processors utilizing different architectures against each other.

(Credit: Michael Justin Allen Sexton)

An important detail about microarchitectures that you should know, and that is easy to understand, is how to identify products based on different architectures. This is surprisingly easy if you learn the numbering systems used by AMD and Intel.

If we take AMD’s Ryzen 9 9950X as an example, for desktop chips the numbers break down like this…

• 9 = Generational number. AMD’s microarchitectures of recent years are named “Zen,” followed by a number. All modern AMD desktop processors that start with a “9” utilize the Zen 5 microarchitecture, and all processors that start with a “7” utilize the Zen 4 microarchitecture. All modern AMD processors that start with a “5” utilize the Zen 3 microarchitecture. AMD has no desktop processors that begin with a “6”; these were mobile only. There were a few desktop Ryzen CPUs that begin with an “8”, and these also run the Zen 4 microarchitecture.

• 9 = Product tier. AMD makes Ryzen 3, 5, 7, and 9 processors, with this numeral included after “Ryzen” in the product name and as the second digit in the product number.

• 5 = Sub-product tier placement.

• 0 = Sub-product tier placement. Rarely used.

• X = “Extreme” or higher-end variant of a processor. An AMD chip may or may not have a non-“X” variant.

(Note that AMD is moving to a wholly different processor numbering scheme for its very latest mobile CPUs. See the details here; the guidelines above will not apply to them.)

Traditionally, Intel used a similar coding scheme for its desktop processors that’s still partly in use. Let’s take an Intel high-end CPU, the Core i9-14900K, as an example…

• 14 = Generational number. Number “14” is assigned to Intel’s 14th Gen “Raptor Lake Refresh” processors. 13 = 13th Gen “Raptor Lake.” 12 = 12th Gen “Alder Lake.” 11 = 11th Gen “Rocket Lake.” 10 = 10th Gen “Comet Lake.”

• 9 = Product tier. Intel makes Core i3, i5, i7, and i9 processors, which are included after the “Core” in the product name and as the third digit in the product number (or the second digit, in processors older than 10th Gen).

• 0 = Sub-product tier placement.

• Second 0 = Sub-product tier placement.

• K = Indicates a higher-end variant of a processor with overclocking support. Several other “suffix” letters are used by Intel to indicate other differences, but the most common ones you will encounter are “K” (indicating overclockability), “F” (indicating that the CPU has no integrated graphics processor), and “KF” (indicating both aspects).

With Intel’s new “Arrow Lake” desktop processors, Intel has adopted a new naming scheme that diverges somewhat from previous generations, but it remains similar. The Core i9, i7, and i5 of previous generations are gone, replaced by Core Ultra 9, Core Ultra 7, and Core Ultra 5 families, respectively. The numbering scheme after these new names has been reduced to three digits, plus optional letter suffixes, which remain largely unchanged from previous generations. Here’s an example of how the new naming scheme works, using the flagship Core Ultra 9 285K chip as an example.

• 2 = Generational number. The number “2” is assigned to Intel’s Arrow Lake processors. No desktop chips were assigned the number “1,” and no additional desktop chips have used this naming sequence yet.

• 8 = Relative performance differentiator. A higher number here general suggests higher performance in relation to other processors in the same product line. All Ultra 9 CPUs have an “8” here, while all Ultra 7 processors have a “6” here. Ultra 5 CPUs can have a “4,” “3” or “2” in this position. This number is not an absolute, however, as ultra-efficiency models like the Ultra 9 285T could potentially be outpaced by chips like the Ultra 7 265K that support higher power ratings.

• 5 = This digit currently holds little meaning; all Arrow Lake desktop processors so far have a “5” as the third digit. We’ll see if that changes down the line.

• K = Indicates a higher-end variant of a processor with overclocking support. Several other “suffix” letters are used by Intel to indicate other differences, but the most common ones you will encounter are “K” (indicating overclockability), “F” (indicating no integrated graphics processor), and “KF” (indicating both aspects). Finally, the “T” suffix is used for desktop processors that are optimized for power efficiency and lower heat output. Some Arrow Lake desktop chips will have no letter at all at the end, terminating just in the “5.”

Using these numbers, you can compare processors of the same generation within a vendor’s own line in terms of relative performance with some accuracy. It’s always best to check reviews when possible and compare other details about the processor to more accurately compare CPUs.

Core Count

Inside all mainstream desktop processors today are multiple CPU cores. In the past, processors only had one CPU core, but as technology has improved, more cores have been manufactured into processor silicon to increase performance. Each CPU core operates as a semi-independent component inside the processor and is capable of completing tasks.

The advantage of having more CPU cores is simply that you can get more work done at a time. It wouldn’t be too much of a stretch to think of the CPU cores as workers and the processor as the building in which they work.

(Credit: Joseph Maldonado)

Traditionally, all of these CPU cores would have been identical, but this changed starting with Intel’s 12th Gen Alder Lake processors. Intel now employs two different types of CPU cores inside most of its desktop processors. Processors based on this design have what Intel calls “P-cores,” which are built on a high-performance microarchitecture. Alongside the “P-cores,” Intel also adds “E-cores,” which utilize a different microarchitecture that enables these cores to be physically smaller and more energy-efficient.

If we return to our comparison of CPU cores to employees, you could think of Intel’s P-cores as higher-level workers with larger offices who can complete more work due to their more extensive experience and larger workspaces. At the same time, the E-cores could be thought of as lower-level workers with smaller offices to work in and lesser skill sets. They may get less done in a given period of time, but you can cram more of them in for the money, and they take up less space.

Though the E-cores are slower, they still significantly improve the processor’s performance. AMD hasn’t adopted a similar scheme yet, which means all of the cores in an AMD processor are identical.

(Credit: Michael Justin Allen Sexton)

Core count contributes greatly to a processor’s overall performance, but this alone does not determine whether one processor is faster than the other. It’s entirely possible for a quad-core processor to be faster than an octa-core processor, and vice versa.

Multithreading

Work that needs to be performed on the CPU cores comes into the processor in a somewhat chaotic fashion. Some processors take the work orders as they come in and simply proceed to work on them in the order they were received. Processors that do this implement what we know as an In-Order execution design. This has been shown to hamper a processor’s overall performance.

Each work order sent to the processor requires instruction information and raw data for the work to be completed. When an In-Order CPU core goes to work on a task, it must have both of these; otherwise, it will sit and wait while the needed instruction information or data is fetched.

Processors that implement an “Out-of-Order” or “OoO” execution design largely circumvent this issue by re-ordering tasks as they come in. They can place work orders that have everything they need to run ahead of work orders that don’t. Inevitably, some work orders still get to the CPU cores without everything they need, though, and this still leads to a stall while the required data gets fetched.

That is, unless a processor implements simultaneous multithreading (SMT, for short). Essentially, what this technology does is open up a second line for work orders to come into the CPU. The processor isn’t able to work on two work orders at once, so when everything is running smoothly, the processor continues to run through work orders from each line in order, back and forth. When a stall occurs, however, SMT technology enables the processor to set the stalled work order aside and work on items in the other line until the required data is fetched.

This technology significantly reduces processor stalls and drastically improves performance. Processors that implement SMT will show as having two threads for each CPU core that supports it. This is why an AMD processor with eight cores and SMT technology will be touted as a 16-thread processor. The same is true for Intel processors, with the most notable difference being that Intel calls SMT “Hyper-Threading” on its processors.

Just like with core count, thread count doesn’t tell you enough to determine which processor is best, but it can give you an idea as to which processor is better in a given line. A processor with more threads may well have a performance advantage over one that has fewer threads supported, in applications that can take advantage of the technology. But as we said with the core count, all of these factors need to be taken into consideration to know for sure.

Clock Speed

The last and easiest-to-understand key defining characteristic of a processor is its clock speed. This directly relates to a processor’s overall speed and is measured in hertz. Processors today are so fast that this is typically reported in gigahertz (GHz).

(Credit: AMD)

A processor’s clock speed is sometimes reported as the total number of operations that it can perform at any given second. For example, a one gigahertz processor can theoretically perform 1,000,000,000 operations each second. Modern CPUs operate at multiple GHz with some, like Intel’s Core i9-14900K, peaking at 6.0GHz.

In truth, this description is inaccurate, as some operations require multiple clock cycles (multiple hertz) to complete, and this is where architecture comes in, coming full circle. When comparing processors that are part of the same generation and product line, it’s safe to think the one with the most cores and the highest clock speed will perform the best. Comparing across different architectures and product lines, however, this is not always the case.

On modern processors, you’ll often see a base clock listed, as well as a “Boost Clock” or “Turbo Clock.” You can essentially ignore the base clock listing if you see either of these other clocks listed. A base clock is a processor’s true baseline speed, but modern CPUs are designed to increase their clock speed to a point, as long as the right conditions are met; those conditions are based on thermal and power-draw limitations.

Modern processors run at these elevated clock speeds most of the time when under a heavy load, which is why these boosted numbers are far more important for determining performance than the base clock. It is also possible to increase the peak clock speed yourself on some processors by what is called overclocking, but that’s for another guide.

To overclock, you’ll need a lot more specific information on the topic, beyond the scope of this article. You’ll also need special hardware for overclocking (an appropriate motherboard, and robust CPU cooling). We point out which processors can overclock in our reviews, and while talking about specific products on this page, so you will know which to buy if you want to try your hand at overclocking. But you should make sure you do plenty of research on the topic first, as overclocking can be hazardous to your components.

How to Buy the Right CPU: Final Buying Advice

Think carefully about your computing needs and pick a processor that fits them. If you have extra cash to spare and want to be extra safe, buy one a little better than your target chip.

One thing we would recommend against is buying the most expensive processor that money can buy or that you can afford without thinking through the decision. Yes, more-expensive processors are typically better, in general—AMD and Intel charge more for them for a reason—but that doesn’t mean you need that level of performance.

No matter how performant a processor you buy, you will eventually want to upgrade to something newer and faster. Though you may be able to prolong the life of your PC by buying a newer and faster CPU, it might make better sense to opt for upgrading more frequently.

The rate at which technology has been improving makes it likely that, in five years or so, you’ll be able to buy a midrange processor that will be just as fast—if not faster than—today’s very fastest processors. If you buy or build a new midrange PC every five years or so, you’ll likely have a faster computer than someone who is still running a five-year-old PC with a high-end processor.

Upgrading more often may not be better financially in the long term, as there is a definite cost and set of diminishing returns involved, but it does get you the improvements that come with changing platforms. As PC components eventually fail, it can also help to avoid unexpected downtime from old parts breaking. For further guidance, check our processor reviews for more details on each CPU.