Intel Core Ultra 5 245K

PCMag editors select and review products independently. If you buy through affiliate links, we may earn commissions, which help support our testing.

Intel has let loose its first volley of new "Arrow Lake" Core Ultra 200S processors, and so far, they've been all over the range. Intel’s Core Ultra 9 285K showed areas of big promise when we tested it, but inconsistencies and lackluster gaming performance also hampered it. Down the stack, Intel’s Core Ultra 5 245K price is certainly more attractive ($309, versus $589 for the 285K flagship chip), and this little processor knows how to stay cool when things get busy. Its gaming performance, however, is worse than the Core Ultra 9 285K's, and while some speed and feature improvements help to make up for this, the Ultra 5 is no gaming champion like the Core i5 chips of yore. It finds itself in a tough-to-recommend position, especially considering the cost to upgraders of moving to a whole new CPU platform. Its cool runnings may make it a good fit for compact, budget-minded PCs, but AMD and 14th Gen Intel alternatives look to be a better value.

The Intel Core Ultra 5 245K on Paper

Before going any further, if you’re interested in the details around the Arrow Lake architecture and what feature improvements Arrow Lake has over the outgoing 14th Gen "Raptor Lake Refresh" processor line, then read our Core Ultra 9 285K review first. A lot has changed with Arrow Lake.

To name a few firsts: These are Intel’s first modern chiplet-based processors, the first produced on a manufacturing process not owned by Intel, and the first desktop Intel processors with dedicated AI silicon. I went over much of this in my Core Ultra 9 285K review, and while I will lightly touch on many of these items here, I won’t do so at nearly the same depth.

Arrow Lake processors, including the Core Ultra 5 245K, are all chiplet designs and use several different manufacturing technologies to make various parts of the processor. This has the potential benefit of improving yields, reducing costs, and increasing production capacity. The most advanced process node used in fabricating Arrow Lake processors is TSMC’s N3B 3nm process, used here exclusively for the Arrow Lake chips' compute tile. This is the fragment of the processor that contains the CPU cores and cache, which makes it likely the most power-hungry and performance-intensive part.

Chips produced on more advanced manufacturing processes typically require less power and produce less heat, so it makes sense to produce the most power-hungry components on the best available process. This manufacturing process is in high demand, however, so making the whole chip on it would be costly and result in fewer chips being made overall. This is why, at least in part, Intel is moving to chiplets and using less advanced manufacturing processes to produce the other portions of the processor.

The Core Ultra 5 245K has six Performance cores (P-cores) based on Intel’s new "Lion Cove" architecture optimized for high performance, and it has eight Efficient cores (E-cores) based on Intel’s "Skymont" architecture optimized for high efficiency. This makes the Core Ultra 5 245K a little better than half of a Core Ultra 9 285K, which has eight P-cores and 16 E-cores. It has an equal number of cores as Intel’s preceding Core i5-14600K, but Arrow Lake doesn’t support Intel’s classic thread-doubling Hyper-Threading technology, which technically gives the older Core i5-14600K an advantage in terms of overall thread count.

However, supported threads aren’t everything, and Arrow Lake has several adjustments that boost its performance. The new E-cores are reportedly up to 32% faster than those found on Raptor Lake processors, and the P-cores are up to 9% faster at the same clock speed. The clock speeds on the Core i5-14600K and the Core Ultra 5 245K aren’t the same, though, and the newer Core Ultra 5 245K has the advantage here.

The Core Ultra 5 245K’s E-cores can hit speeds up to 4.6GHz, a considerable increase from the 4.0GHz clock ceiling of the Core i5-14600K’s E-cores. At the same time, the Core Ultra 5 245K’s P-cores are clocked a little lower (5.2GHz) than the Core i5-14600K’s P-cores (5.3GHz) at boost, but the performance improvements alone should help to make up for this, with the E-cores helping to push performance up even further.

Intel also equipped the Core Ultra 5 245K with an upgraded integrated graphics processor (IGP) based on its Xe-LPG architecture. This is the same architecture used inside Intel’s current Arc graphics cards, but it’s been altered slightly to transform it into an integrated solution fused into the processor. I couldn’t test this IGP as the current motherboard on our testbed lacks any video outputs, but the new IGP has higher clock speeds, an increase in execution units from 32 to 64, and ray-tracing hardware. All in all, this should provide a significant performance boost.

The memory controller on Arrow Lake processors was also improved. It now officially supports DDR5 at 6,400MHz, up from 5,600MHz on Raptor Lake processors. (The Z890 platform's support for optional CU-DIMM memory modules, equipped with an on-module clock generator, can also enable higher memory speeds.)

Testing the Intel Core Ultra 5 245K: Setup & Competition

Our Core Ultra 5 245K sample was tested on an MSI MEG Z890 Unify-X motherboard with two 16GB sticks of DDR5 RAM clocked at the processor’s max officially supported speed of 6,400MHz. The processor was cooled by a Cooler Master MasterLiquid PL240 Flux 240mm water cooler, and a SilverStone DA850 850-watt power supply powered the system. An Nvidia GeForce RTX 4080 Super Founders Edition card was the display adapter for all gaming tests. The whole system was built onto a Praxis Wetbench open-air testbed with a 1TB PCIe 4.0 NVMe SSD with Windows 11 installed.

Arriving in the middle of the processor market, the Core Ultra 5 245K has plenty of competition to face off against. Its $309 price fits directly between AMD’s Ryzen 5 9600X, which is $279, and the Ryzen 7 9700X, which is $359. Intel also has the Core i5-14600K, which is priced at $319.

Testing the Intel Core Ultra 5 245K: CPU Performance

On paper, Arrow Lake looks like it should impress. As always, we rely on benchmarks to help us determine how effective a processor design really is, and as it turns out, the benchmarks here give a mixed picture. Naturally, we would expect to see the Ultra 5 245K beat out its predecessor in most ways, but that's not how it shook out.

Starting with Cinebench 2024, the Ultra 5 245K drew a tie with the Core i5-14600K in the multi-threaded test. In a way, this shows the added power of the Arrow Lake architecture, as the 245K lacks Hyper-Threading support, which undoubtedly helped the Core i5-14600K. That the 245K’s 24 CPU cores and 24 threads matched the 14600K’s 24 CPU cores and 30 threads is an accomplishment. But technical victories alone don’t usually persuade people to upgrade.

On a more positive note, the Ultra 5 245K did hold a safe lead over the Ryzen 7 9700X and the Ryzen 5 9600X in the multi-threaded Cinebench 2024 test. Single-threaded performance results showed the opposite, with Ultra 5 245K slightly behind its two AMD competitors and significantly ahead of its Core i5 predecessor.

Blender, likewise, produced somewhat mixed results. On the plus side, the Ultra 5 245K was again far faster than the Ryzen 5 and Ryzen 7 competition. Compared with the Ryzen 7 9700X, the Ultra 5 245K was 22% faster on the benchmark's Monster and Classroom sample projects, with an 11% lead in the Junkshop one. These performance leads grow more significant if we calculate with the Ryzen 5 9600X instead of the Ryzen 7 9700X. The Core i5-14600K was a closer match for the Ultra 5 245K, with slightly faster performance in Junkshop and slightly lower performance in the other two Blender tests. But it wasn’t a clear win in that Intel-versus-Intel comparison for the Ultra 5 245K.

HandBrake was a more decisive test, with the Ultra 5 245K holding a clearer lead. It was just 5% faster than the Core i5-14600K here, but a win is a win. At the same time, the Ultra 5 245K was 29% faster than the Ryzen 7 9700X and 55% faster than the Ryzen 5 9600X. This trend continued in POV-Ray; the Ultra 5 245K again had an unambiguous win, completing the multi-threaded test in 12.5% less time than the Core i5-14600K, 30% less time than the Ryzen 7 9700X and 42% less time than the Ryzen 5 9600X.

With AIDA64, we suspect software-optimization issues caused the Arrow Lake and Ryzen 9000-series processors to perform significantly worse in the AES 256-bit and SHA-1 Hash tests. Though it’s possible the Raptor Lake processors could simply be far better at these specific workloads, it just doesn’t quite match up with the indications we get from all the other tests, including the AIDA64 floating-point tests. In these, we saw the Ultra 5 245K and the Ryzen 5 9600X in a near tie, surpassing the Core i5-14600K by a wide margin and lagging far behind the Ryzen 7 9700X.

Adobe’s Photoshop test also suggests that the software needs an update to work better with Arrow Lake processors. In this case, the Ryzen 9000-series processors performed well and seemed to work correctly, but it doesn’t make much sense for the Core Ultra 5 245K to beat out the Core Ultra 9 285K as it did here. As for the Adobe Premiere Pro test, we saw fewer issues, with the Ultra 5 245K performing similarly to the Ryzen 7 9700X but slower than the Core i5-14600K.

Testing the Intel Core Ultra 5 245K: Gaming Performance

Intel’s Core i5 processors have long been regarded as excellent sweet-spot gaming processors. As the successor to those venerables, the Ultra 5 245K has big shoes to fill and was expected to be a natural go-to option for gaming PCs. That's not how our gaming benchmarks played out, however.

Out of five games tested at three resolutions (a total of 15 separate tests), the number of tests in which the Core Ultra 5 245K beat the Core i5-14600K is...zero. You'll find a few tests in which the two processors tied, but out of the 15, the Core i5-14600K was faster in 12. If we count performance differences of 3% or less as a tie, which we typically do, this count dropped to eight of 15, but where it prevailed, the Core i5-14600K was sometimes up to 17% faster. No matter how you look at it or count it, the Ultra 5 245K looks to be a lesser gaming option than its predecessor.

We see similar results compared with the AMD competition. Here again, the Ultra 5 245K, at best, tied in with the Ryzen 5 9600X and the Ryzen 7 9700X, and at worst, it fell up to 9% behind the Ryzen 5 9600X.

Saying that the Intel Core Ultra 5 245K is a poor CPU for gaming is a stretch, as it still performed reasonably well overall. It didn’t hinder frame rates by so much as to make anything unplayable, and most games ran fast enough that you wouldn’t notice a whole lot of difference between the Ultra 5 245K and its competition. (And we saw quite a few ties.) But given that this is a new platform that will require the expense of a new Z890 motherboard, you also can't deny that all of the competitors listed here, from both AMD and Intel, are better options for gaming, at least for now, especially if you already have a compatible board.

Testing the Core Ultra 5 245K: Power Consumption & Thermals

We use a Kill-A-Watt power meter to gauge the Ultra 5 245K's power consumption and that of other tested CPUs by checking the meter during specific tests. Power-consumption numbers may be a little off between AMD and Intel systems due to the unavoidable differences in the motherboards, but in broad strokes, this should give us a sensible indication of how much power each processor draws.

We see the Ultra 5 245K’s best qualities here. Topping out at around 205 watts, the Core Ultra 5 245K is one of the less power-hungry CPUs Intel has released in recent years. At load, it consumed about 26% less power than the Core i5-14600K, but it also pulled a fair bit more power than the Ryzen 7 9700X and the Ryzen 5 9600X. Both of those AMD CPUs have lower (65-watt) TDPs, which helps keep them from being excessively power-hungry.

The power-draw numbers suggest the Ultra 5 245K is a fair bit more energy efficient than its predecessor, but it’s harder to say which is more efficient, the AMD CPUs or the Core Ultra 5 245K. The lower power figures suggest that the Ryzen CPUs would be more efficient, but Intel’s higher performance results in many CPU tests blunt that advantage.

For example, while the Ryzen 7 9700X used 22% less power in Cinebench 2024 than the Ultra 5 245K, the Ultra 5 scored 23% higher in performance during the test. This makes the two CPUs, in essence, similar in terms of efficiency for this test. The truth is: Which processor is more efficient will depend upon which program it's operating under.

Operating temperatures on the Core Ultra 5 245K were exceptional, with the processor never exceeding 68 degrees Celsius during the monitored portion of our testing. AMD’s Ryzen 5 9600X and Ryzen 7 9700X performed much the same here, though the Ultra 5 245K held a slight edge. The older Core i5-14600K ran far hotter, which again signals a serious improvement here from Intel.

Verdict: Arrow Lake Hits the Target, Just the Wrong One

With such a jagged landscape making up the 245K's performance tests, judging this chip is difficult. We see some obvious improvements here over 14th Gen Core, that much is certain. But are those improvements in the right areas? Not so much.

After what we saw with Raptor Lake and Raptor Lake Refresh, it was clear to all that Intel needed to stem its ever-escalating power consumption and heat production. Intel has significantly improved on both fronts, with the two Arrow Lake processors we’ve tested so far showing better power-consumption figures, and the Ultra 5 245K showing substantially lower operating temperatures. After a season of unstable processors in Intel's 13th and 14th Gen, this was arguably the most critical thing for Intel to do, and it succeeded.

Performance is also up in select areas by a considerable amount, but at the same time, it’s down in others. This is particularly true for gaming workloads, where the Ultra 5 245K failed to stand up to the competition.

The biggest thorn in the Ultra 5 245K’s side, though, is Intel's own Core i5-14600K, which has an MSRP of $319 but retails for closer to $259 today. In other words, the 14600K isn’t just better for gaming than the Ultra 5 245K, but it also costs less. The gains that the Ultra 5 245K holds aren’t zilch, but they just aren’t in the places where most customers will care enough to make the leap, and the performance improvements are simply not consistent enough.

Until Intel pushes some sort of patch to help resolve some of these issues (that is, assuming it can and does), opting for the older Core i5-14600K, AMD’s Ryzen 5 9600X, or AMD’s Ryzen 7 9700X over the Ultra 5 245K will make a lot more value sense, particularly if you are building a gaming PC.