Today we are looking at Intel's last decade of flagship processors, touching on what each series brought to the table. We will cover the history of each generation and then show how it performs today across a range of modern games, giving us a clear view of where and how Intel has made gains over the years. It is an interesting journey, and we think you're going to enjoy this one, so let's get into it.
The 7th generation: Kaby Lake, featuring the Core i7-7700K
Released back in 2017, we begin with the Core i7-7700K, a refresh of the 6700K. In other words, this is CPU technology that first appeared in August 2015. Both chips are nearly identical, featuring a 4 core, 8 thread design with 8MB of L3 cache and a 91W TDP.
While the CPU cores remained unchanged from Skylake, the integrated GPU received a meaningful, if somewhat niche, upgrade. The media engine was significantly improved, adding fixed function hardware decoding for HEVC 10-bit and VP9. At the time, this was a major development because it was the only PC hardware that officially supported 4K Netflix streaming on Windows, making it especially appealing for high-end home theater PCs.
For gaming, however, there was little to get excited about. Core clock frequency increased by 7%, resulting in roughly 5% better performance on average, which was underwhelming.
Intel leaned on its highly optimized 14nm+ process to push the 7700K to a 4.5GHz boost clock, with most chips capable of reaching 5GHz when overclocked. By this point, the silicon itself was no longer the primary limitation for enthusiasts. Instead, the bottleneck was Intel's inexpensive thermal interface material, widely mocked by the community as "toothpaste."
Rather than soldering the integrated heat spreader (IHS) to the silicon die, Intel used thermal paste, which is far less effective at transferring heat. This led many enthusiasts to purchase delidding tools to remove the heat spreader, replace the stock paste with liquid metal, and significantly improve thermals. Doing so often eliminated thermal throttling and made sustained 5GHz operation possible within safe temperature limits.
Even so, despite the enthusiasm around delidding, the 7700K and the broader Kaby Lake lineup were ultimately underwhelming. This was especially true given that, at a similar price point, AMD was preparing to introduce 8 core, 16 thread Zen processors.
Kaby Lake marked the end of Intel's undisputed quad-core dominance.
It also meant that after 18 months, Intel's offering for consumers was effectively a weak refresh on the LGA 1151 platform, and that refresh marked the end of the line for the socket. In hindsight, the most notable aspect of the 7700K and Kaby Lake is that they marked the end of Intel's undisputed quad core dominance.
As much as we have criticized the 7700K over the years, it was not a bad processor. It was simply poorly timed, arriving just before the much stronger 8700K later that same year. For those who did buy one, it delivered solid performance, and even today, nearly a decade later, it remains usable in many modern titles.
Across our 14 game test suite, it delivered an average of 78 fps, with 1% lows averaging 54 fps. This is not impressive by modern standards, but it is still generally playable. Performance varies by title. In games such as Counter Strike 2 and Rainbow Six Siege, it is possible to maintain over 100 fps, though most other titles see 1% lows dip below 60 fps.
The 7700K struggles significantly with ray tracing enabled, as seen in Cyberpunk 2077 and Spider Man 2, both tested with ray tracing active. This is largely because the CPU is responsible for building and updating the bounding volume hierarchy, or BVH, a data structure that tracks how rays intersect with objects in a scene.
Ray tracing dramatically increases the amount of data the CPU must process compared to traditional rasterization. In rasterized rendering, aggressive culling reduces workload, as the GPU only processes what is visible to the camera. Objects behind the player are often ignored entirely. With ray tracing, that assumption no longer holds. Objects outside the field of view can still influence reflections and lighting, meaning the CPU must process and send far more geometry data to the GPU.
In the case of the 7700K, which is already operating near its limits, this additional workload causes performance to drop sharply. This is why we see sub 60 fps results in Cyberpunk 2077 and Spider Man 2. That said, ray tracing is not the only factor. Any CPU intensive workload can expose these limitations, as seen in titles like Mafia: The Old Country.
In most games, disabling hardware based ray tracing can significantly improve performance. When we do this in Cyberpunk 2077 and Spider Man 2 using the medium preset, performance recovers considerably. Average frame rates climb into the 90s, with 1% lows exceeding 60 fps. It is worth noting that this preset also reduces other CPU intensive settings such as crowd density, but ray tracing remains one of the most demanding features for the CPU.
The 8th generation: Coffee Lake, featuring the Core i7-8700K
Fast forward just nine months, and Intel was already feeling pressure from AMD's newly released Zen processors, which offered twice as many cores as Intel's best mainstream parts. The company needed to respond, and this is when the LGA 1151 socket trap was sprung.
In typical Intel fashion, the response was to introduce yet another new socket that was incompatible with the previous one. What made this move particularly frustrating was that it was physically the same socket, LGA 1151. Intel simply labeled it as a second revision and, despite being pin compatible, blocked support at the BIOS level. This meant you could not install the Core i7-8700K on an existing LGA 1151 motherboard. Instead, you needed an LGA 1151 revision 2 board. In many ways, this strategy played a role in the decline that followed in the years ahead.
Intel reused the same LGA 1151 socket, then blocked compatibility at the BIOS level.
Not only was the socket physically the same, so was the CPU architecture. Codenamed Coffee Lake, the 8th generation Core series was not based on a new microarchitecture, nor did it introduce a smaller manufacturing node. At the time, Intel was facing significant delays with its 10nm process, so the 8700K was built on a further refinement of the existing 14nm node, officially called 14nm++. This refinement increased transistor density and power headroom, allowing for higher clock speeds out of the box without changing the physical package.
8th Gen Intel Core Processor Family
| Core i3 8100 | Core i3 8350K | Core i5 8400 | Core i5 8600K | Core i7 8700 | Core i7 8700K | |
| Cores / Threads | 4 / 4 | 4 / 4 | 6 / 6 | 6 / 6 | 6 / 12 | 6 / 12 |
| Base Frequency | 3.6 GHz | 4.0 GHz | 2.8 GHz | 3.6 GHz | 3.2 GHz | 3.7 GHz |
| Max Turbo Frequency | N/A | N/A | 4.0 GHz | 4.3 GHz | 4.6 GHz | 4.7 GHz |
| L2 Cache | 1 MB | 1.5 MB | ||||
| L3 Cache | 6 MB | 8 MB | 9 MB | 12 MB | ||
| iGPU Model | UHD Graphics 630 | |||||
| PCIe Lanes | 16 PCI Express 3.0 Lanes | |||||
| Memory Config | Dual-Channel | |||||
| Max Mem Support | DDR4-2400 | DDR4-2666 | ||||
| Socket | LGA 1151v2 | |||||
| TDP | 65 W | 91 W | 65 W | 95 W | 65 W | 95 W |
| MSRP | $117 | $168 | $182 | $257 | $303 | $359 |
Because Coffee Lake was an optimization rather than a new architecture, instructions per clock remained identical to Skylake and Kaby Lake. Any single threaded gains seen with the 8700K over the 7700K came from higher clock speeds rather than improved efficiency. The real upgrade was the addition of two extra cores, resulting in a 6 core, 12 thread configuration with 12 MB of L3 cache.
The jump to 6 cores and 12 threads was a turning point for Intel's mainstream CPUs.
These changes delivered a substantial performance boost, although the real gains, especially in gaming, were not immediately realized. At the time, most games were still designed with quad core CPUs in mind. Today, however, the gains are clear.
While the 8700K still struggles in titles such as Spider Man 2 with ray tracing enabled, it is 36% faster than the 7700K, which represents a massive generational leap.
We're looking at an even larger 49% increase in Cyberpunk which is right in line with the 50% increase in cores, though there are other factors at play here like core clock speed and cache capacity.
The big improvements are not universal though. In Mafia: The Old Country, for example, the 8700K was just shy of 20% faster. Even so, it is clear that the 8700K is a far more capable processor. With reduced quality settings, such as the medium preset used here, it remains very usable today and can deliver high refresh rate performance in a number of modern titles.
The 9th generation: Coffee Lake Refresh, featuring the Core i9-9900K
The Core i7-8700K was a great processor, even if it frustrated existing Intel users on Skylake or Kaby Lake platforms. Still, just a year later, Intel was once again under pressure from AMD. With its 10nm process still delayed, the company returned to 14nm++, this time adding two more cores and introducing updated branding.
Before the 9th generation, the Core i9 name was reserved for Intel's HEDT high-end desktop platform, such as the Skylake X series. The Core i9-9900K brought that premium branding to the mainstream for the first time. It featured 8 cores and 16 threads, allowing Intel to finally match AMD's Ryzen 7 lineup in core count.
Intel also addressed the criticism over its use of thermal paste between the silicon die and the integrated heat spreader. With the 9900K, the company reintroduced solder thermal interface material, or STIM. In practice, this was likely a necessity. Packing 8 high frequency cores onto the 14nm node generated significant heat, and solder provided a more efficient path for transferring that heat to the heat spreader, helping to prevent immediate thermal throttling under heavy multi-core workloads.
That said, thermal behavior remained a concern. Intel officially rated the i9-9900K with a 95 W thermal design power, but this figure applied only to the base clock of 3.6 GHz. Under sustained multi threaded workloads such as rendering, the processor could draw around 200W of package power. This created problems for users pairing the 9900K with entry level or mid-range Z390 motherboards, many of which used weaker VRM designs that could overheat and limit performance.
Despite these issues, Intel retained the gaming performance crown with the 9900K. Even today, while it is clearly showing its age, it remains a capable processor that can still deliver solid performance in modern games.
The 10th generation: Comet Lake, featuring the Core i9-10900K
It was clear that the 9900K and the broader Coffee Lake lineup would not be enough to hold back AMD's Ryzen momentum. Just six months later, Intel pushed out its 10th generation series on the new LGA 1200 socket.
This brought the Core i9-10900K, marking the twilight of Intel's 14nm era. Facing intense pressure from AMD's Ryzen 3000 series, which scaled up to 16 cores, Intel was forced to push its aging Skylake-derived architecture to unprecedented extremes.
The 10900K was the final stretch of Intel's 14nm era pushed to its limits.
Comet Lake-S was yet another refinement of the 14nm process, often referred to as 14nm+++ by the community. By this point, Intel had been relying on variations of the same node for five years, dating back to its introduction with the 5th generation Broadwell architecture in mid-2015.
To Intel's credit, the company managed to cram 10 cores and 20 threads onto a mainstream die. However, because it was still using the Skylake core design and ring bus interconnect from 2015, the silicon footprint had grown significantly. Scaling beyond 10 cores would likely have required removing the integrated graphics entirely just to make the design viable.
Managing thermals became a major challenge, so in order to handle the immense heat generated by ten 14nm cores running at high frequencies, Intel had to get creative with the physical packaging. To improve heat dissipation, Intel reduced the thickness of the silicon die by 0.3 mm, shortening the distance heat needed to travel. To maintain the same overall package height, the copper integrated heat spreader was made 0.3 mm thicker.
In an effort to reclaim the "world's fastest gaming processor" title, Intel introduced a new boosting mechanism called Thermal Velocity Boost. The 10900K featured a base clock of 3.7GHz and a standard turbo of 5.1GHz, while Thermal Velocity Boost allowed single-core boosts up to 5.3GHz. However, this peak frequency was only achievable if CPU temperatures remained below 70°C. Once that threshold was exceeded, clock speeds would drop.
Perhaps the most frustrating aspect of Comet Lake-S was not the requirement for a new 400 series motherboard, such as Z490, as that had become standard practice for Intel. Rather, it was the fact that many of these boards included PCI Express 4.0 support at the hardware level, yet the feature remained disabled until users upgraded to 11th generation processors a year later. This stood in contrast to AMD platforms, which were already taking advantage of PCIe 4.0 for high-speed NVMe storage.
Even so, the addition of two more cores and increased L3 cache did result in improved performance. The 10900K was faster than the 9900K, although the gains were relatively modest. At this point, it was clear Intel needed a more fundamental shift, but the underlying technology was not yet ready to deliver it.
The 11th generation: Rocket Lake, featuring the Core i9-11900K
Out of options, Intel moved forward with its PCIe 4.0 enabled processors for the LGA 1200 platform, codenamed Rocket Lake. The problem was that the company remained stuck on the 14nm process. What followed became one of the most unusual and heavily debated chapters in Intel's history. Rocket Lake was a stopgap, a brute force attempt to maintain a performance lead while Intel struggled to bring its delayed 10nm manufacturing process online.
Rather than continuing to recycle the Skylake core yet again, Intel took a far more ambitious approach. The company adapted its Sunny Cove architecture, originally designed for 10nm mobile chips, and brought it over to the 14nm process. This backporting effort became known as Cypress Cove. It was an impressive engineering achievement, but it came with significant tradeoffs that ultimately hurt both performance and market reception.
Rocket Lake was a stopgap that traded efficiency for complexity and lost ground in the process.
Adapting a 10nm design to 14nm resulted in much larger cores. The increased size quickly became a limiting factor, as Intel ran out of space on the die. To fit within the constraints of the LGA 1200 package while retaining integrated graphics, the company had to reduce the core count. As a result, the flagship Core i9-11900K launched with 8 cores and 16 threads, a step back from the 10-core configuration of the previous generation's 10900K.
Intel Core i9-11900K Specs
| Core i9-11900K | Core i9-10900K | Core i9-10900KF | Ryzen 9 5900X | |
| MSRP | $539 | $488 | $472 | $549 |
| Current Price | $615 | $460 | $420 | $550 |
| Cores / Threads | 8 / 16 | 10 / 20 | 12 / 24 | |
| Base Frequency | 3.5 GHz | 3.7 GHz | 3.7 GHz | |
| Max Turbo | 5.3 GHz | 5.3 GHz | 4.8 GHz | |
| L3 Cache | 16 MB | 20 MB | 64 MB | |
| iGPU Model | UHD Graphics 750 | UHD Graphics 630 | N/A | |
| TDP | 125 watt | 105 watt | ||
After several years of development, this left Intel in a difficult position. Gaming performance had effectively peaked with the 10th generation in 2020, following multiple generations of essentially tacking on more cores and cache, while pushing frequencies and power draw to the moon.
Still in modern game, Intel delivered roughly a 55% performance improvement over a three year span, before regressing slightly with the 11900K. By this point, AMD had caught up and moved ahead with its Zen 3 architecture, achieving approximately a 65% improvement over the same period.
The 12th generation: Alder Lake, featuring the Core i9-12900K
Perhaps the most astonishing part of this timeline is that just eight months after the disastrous 11th generation, Intel delivered a radically redesigned 12th-gen – on a new socket, of course.
The Core i9-12900K and the Alder Lake architecture represented the most significant redesign of Intel's desktop processors in over a decade. After years of incremental updates on 14nm and losing ground to AMD, Alder Lake was a major technological leap that put Intel back in a strong position.
Alder Lake was Intel's biggest desktop CPU redesign in over a decade.
After being stuck on 14nm since 2015, Alder Lake finally moved things forward. The 12900K was built on a new node branded as "Intel 7." Despite the name, this was not a true 7nm process, but rather a refined version of Intel's 10nm Enhanced SuperFin node. The new naming reflected competitive transistor density and performance relative to other 7nm class processes, such as those produced by TSMC.
The architectural changes were even more significant. For decades, mainstream desktop CPUs used a homogeneous design where all cores were identical. Alder Lake threw this out the window by adopting a hybrid approach similar to what had long been used by the likes of Arm in smartphone processors.
The 12900K featured 8 performance cores or P-Cores for heavy, latency sensitive workloads like gaming, alongside 8 efficient cores or E-Cores designed for background tasks and additional multi-threaded throughput. Because the efficient cores do not support Hyper-Threading, the processor used a unique 16-core, 24-thread configuration.
Intel's 12th generation also introduced DDR5 memory and PCI Express 5.0 to the mainstream desktop. At launch, however, DDR5 offered little real world benefit. In our day-one testing, the 12900K was up to 2% faster using the much higher bandwidth DDR5 compared to DDR4. The newer memory was also extremely expensive at the time, much like today.
As with the transition beyond quad-core CPUs, software needed time to catch up. When the 12900K launched, most games were not designed to take advantage of the additional bandwidth, starving DDR5 of the opportunity to shine. Over time, that has changed. Modern games now make far greater use of memory bandwidth, particularly with more advanced rendering techniques such as ray tracing.
With DDR4, the 12900K was barely faster than the 11900K.
Looking back at our original review data, the 12900K was on average 8% faster than the 11900K when using DDR4 memory. In updated testing with faster DDR4, that margin narrows to around 4%. In practical terms, the difference remains small. The point was that using DDR4 memory the 12900K really wasn't much faster than the 11900K back in 2021, and that remains true to this day.
But if we arm the 12900K with DDR5 memory, the picture changes significantly. The 12900K is now 23% faster than the 11900K, representing the kind of generational uplift expected from a major architectural shift. Even when limited to DDR4, the 12900K delivers a 59% improvement over the 7700K, which is solid progress over a four year span. That said, this performance comes with a much larger die, which has grown by roughly 76%.
The 13th and 14th generation: Raptor Lake, featuring the Core i9-13900K
Next, Intel introduced the Raptor Lake architecture, a refinement of Alder Lake, with the Core i9-13900K leading the lineup. The key improvements focused on the memory subsystem. To keep the performance cores fed with data and reduce latency, which is critical for gaming, Intel significantly increased L2 cache capacity.
The L2 cache per P-core grew from 1.25MB to 2MB, while the L2 cache per E-core cluster doubled from 2MB to 4MB. This substantial increase in fast on-die memory was a direct counter-strategy to AMD's 3D V-Cache strategy.
Intel Core i9-13900K vs i9-12900K
| Intel Core i9-13900K | Intel Core i9-12900K | |
| MSRP | $600 | |
| Release Date | October 20th, 2022 | November 4th, 2021 |
| P-Cores / E-Cores | 8c [16t] / 16c [16t] | 8c [16t] / 8c [8t] |
| P-Core Clock, Base/Turbo | 3.0 GHz / 5.8 GHz | 3.2 GHz / 5.2 GHz |
| E-Core Clock, Base/Turbo | 2.2 GHz / 4.3 GHz | 2.4 GHz / 3.9 GHz |
| P-Cores / E-Cores, L2 Cache | 8 × 2 MB / 4 × 4 MB | 8 × 1.25 MB / 2 × 2 MB |
| L3 Cache | 36 MB | 30 MB |
| Base Power | 125 watts | 125 watts |
| Max Turbo | 253 watts | 241 watts |
| Socket | LGA 1700 | |
While the 13900K retained the same 8 P-cores as its predecessor, Intel focused heavily on scaling multi-core performance by doubling the number of E-cores. The result was a 24-core, 32-thread processor capable of competing directly with AMD's 16-core Ryzen 9 7950X in demanding workloads such as rendering and video encoding, shifting the competitive landscape.
Raptor Lake was built on a further refined version of the Intel 7 node, a mature 10nm class process, which enabled extreme clock speeds. Out of the box, the 13900K could boost up to 5.8GHz on its P-cores, a massive 600MHz jump over the 12900K. The silicon was so highly tuned that it was pushed even further with the release of the special edition i9-13900KS months later. That became the first consumer processor to officially hit the 6.0 GHz barrier without manual overclocking.
Intel Core i9-13900KS Specifications
| Intel Core i9-13900KS | Intel Core i9-13900K | |
| MSRP | $699 | $600 |
| Release Date | January 2023 | October 2022 |
| P-Cores / E-Cores | 8c [16t] / 16c [16t] | |
| P-Core Clock, Base/Turbo | 3.2 GHz / 6.0 GHz | 3.0 GHz / 5.8 GHz |
| E-Core Clock, Base/Turbo | 2.2 GHz / 4.3 GHz | |
| P-Cores / E-Cores, L2 Cache | 8 × 2 MB / 4 × 4 MB | |
| L3 Cache | 24 MB | |
| Base Power | 150 watts | 125 watts |
| Max Turbo | 253 watts | |
| Socket | LGA 1700 | |
Of course, with great power comes great responsibility, or in the case of the 13900K great power consumption. Intel rated the processor with a maximum turbo power of 253W, but in practice, many motherboards ignored these limits. Under heavy all-core workloads with power restrictions removed, the 13900K could draw between 300W and 350W.
The 13900K pushed desktop CPU power consumption to new extremes.
As a result, Intel shifted its messaging around thermals. They explicitly stated that hitting 100°C under heavy loads was "operating as intended" and completely safe for the silicon. The processor was engineered to use all available thermal headroom, often requiring high-end 360mm or 420mm liquid cooling solutions to maintain peak boost performance.
Sadly, after a few years on the market reports of stability issues for Intel's 13th and 14th gen Core i9 and Core i7 processors started to surface, eventually linked to degradation issues.
High voltage spikes led to accelerated silicon degradation in 13th and 14th gen CPUs.
After months of investigation, Intel identified the primary root cause as Vmin Shift Instability. A flaw in the CPU's microcode algorithm was requesting incorrectly high voltage levels during certain workloads. When combined with the high clock speeds and temperatures of the 13900K, these sustained high-voltage spikes caused accelerated physical degradation of the silicon. As the silicon degraded, the minimum voltage required to keep the chip stable at its rated frequencies shifted higher, creating a feedback loop of instability.
Acknowledging the severity of the issue, Intel extended the warranty on boxed 13th and 14th-gen Core processors (including the 13900K) by an additional two years, giving affected users more time to replace damaged chips.
Despite these issues, the 13900K delivered strong performance, as did the refreshed 14900K. However, both were difficult to recommend in practice due to their power consumption and thermal demands, with the degradation concerns ultimately overshadowing the series.
According to our data, the 13900K is on average 15% to 16% faster than the 12900K in gaming, an impressive uplift, although much of that gain comes at the cost of significantly higher power usage. As for the 14th generation, it is largely a minor refresh. The 14900K, for example, is effectively a slightly higher clocked version of the 13900K, offering around a 3% increase in frequency, similar in concept to the 13900KS.
The Core Ultra 200 series: Arrow Lake, featuring the Core Ultra 9 285K
That brings us to the present day with Intel's Core Ultra 200 series. The Arrow Lake architecture represents a significant philosophical shift for Intel. After years of pushing aging designs to their thermal and power limits, the company has reworked its approach, focusing on power efficiency and a modular chip layout.
For the first time in over a decade, Intel removed simultaneous multithreading, previously known as Hyper-Threading. The Lion Cove performance cores in the 285K are designed for single-thread efficiency and reduced physical footprint, meaning they no longer handle two threads per core. As a result, the 285K features 24 physical cores, consisting of 8 P-cores and 16 E-cores, and a total of 24 threads.
| Core Ultra Model | 285K | 265K | 245K | i9-14900K | i7-14700K | i5-14600K |
|---|---|---|---|---|---|---|
| Release date | Oct 2024 | Oct 2023 | ||||
| MSRP | $600 | $395 | $310 | $600 | $420 | $330 |
| P-core (performance) | ||||||
| Cores (threads) | 8 (8) | 6 (6) | 8 (16) | 6 (12) | ||
| Frequency | 3.7 GHz | 3.9 GHz | 4.2 GHz | 3.2 GHz | 3.4 GHz | 3.5 GHz |
| Turbo | 5.6 GHz | 5.5 GHz | 5.2 GHz | 6.0 GHz | 5.6 GHz | 5.3 GHz |
| L2 cache | 24 MB | 18 MB | 16 MB | 12 MB | ||
| E-core (efficiency) | ||||||
| Cores (threads) | 16 (16) | 12 (12) | 8 (8) | 16 (16) | 12 (12) | 8 (8) |
| Frequency | 3.2 GHz | 3.3 GHz | 3.6 GHz | 2.4 GHz | 2.5 GHz | 2.6 GHz |
| Turbo | 4.6 GHz | 4.4 GHz | 4.3 GHz | 4.0 GHz | ||
| L2 cache | 16 MB | 12 MB | 8 MB | 16 MB | 12 MB | 8 MB |
| L3 cache | 36 MB | 30 MB | 24 MB | 36 MB | 33 MB | 24 MB |
| Power | ||||||
| Base | 125 W | |||||
| Max Turbo | 250 W | 159 W | 253 W | 181 W | ||
Arrow Lake marked a shift from peak performance to efficiency and modular design.
Not only that, but Arrow Lake also marks a major shift in manufacturing strategy. Historically, Intel's advantage came from owning and operating its own foundries, allowing them to manufacture their own chips. Arrow Lake broke this tradition in a major way. To achieve their efficiency goals, the most critical part of the 285K CPU – the "Compute Tile" which houses the CPU cores, was fabricated by rival foundry TSMC, using its cutting-edge 3nm N3B process.
Intel has also abandoned the traditional monolithic die design they had used for generations on the desktop. Instead, Arrow Lake uses a disaggregated multi-chip module approach. The processor is built from several distinct silicon tiles, including a compute tile, a graphics tile, an SoC tile, and an I/O tile, all connected using Intel's Foveros 3D packaging technology.
There is, however, a consequence to this new direction, especially when viewed in the context of the previous few generations. By pushing frequency and power so aggressively in parts like the 13900K and 14900K, Intel effectively set expectations around raw clock speed that Arrow Lake no longer tries to match. The 285K takes a noticeable step back here, with P-cores peaking at 5.7GHz, compared to the 6.2GHz seen with the 14900K. This is a deliberate trade-off, accepting lower peak frequency in exchange for the efficiency gains offered by a denser 3nm class process.
This clock speed regression, combined with the physical separation of tiles and the removal of Hyper-Threading, introduces higher core-to-core and memory latency. This ends up hurting gaming performance more than expected. Even with support for CUDIMM memory pushing frequencies higher, the gaming performance of parts like the 285K has been underwhelming.
For example, in our most up to date gaming data, the 285K is slower than both the 13900K and 14900K. In fact, it is only around 5% to 7% faster than the 12900K, and these are the same margins we observed at launch. In other words, performance has not meaningfully improved since release.
The Core Ultra Plus series and Nova Lake future
More recently, Intel released updated 200 series processors with the 250K Plus and 270K Plus, both of which are notably better than the original models. Not only are they configured more effectively, delivering stronger performance, but Intel has also reduced pricing to better reflect its current position in the market, which is now very much that of the underdog.
Intel Core Ultra 200 Series
| Core Ultra Model | 285K | 270K Plus | 265K | 250K Plus | 245K | 225 |
|---|---|---|---|---|---|---|
| Release date | Oct 2024 | Mar 2026 | Oct 2024 | Mar 2026 | Oct 2024 | Jan 2025 |
| MSRP | $600 | $300 | $395 | $200 | $310 | $240 |
| P-core (performance) | ||||||
| Cores (threads) | 8 (8) | 6 (6) | 6 (6) | |||
| Frequency | 3.7 GHz | 3.9 GHz | 4.2 GHz | 3.3 GHz | ||
| Turbo | 5.6 GHz | 5.5 GHz | 5.3 GHz | 5.2 GHz | 4.9 GHz | |
| L2 cache | 24 MB | 18 MB | 18 MB | |||
| E-core (efficiency) | ||||||
| Cores (threads) | 16 (16) | 12 (12) | 8 (8) | 4 (4) | ||
| Frequency | 3.2 GHz | 3.3 GHz | 3.6 GHz | 2.7 GHz | ||
| Turbo | 4.6 GHz | 4.6 GHz | 4.6 GHz | 4.4 GHz | ||
| L2 cache | 16 MB | 12 MB | 8 MB | 4 MB | ||
| L3 cache | 36 MB | 30 MB | 24 MB | 20 MB | ||
| Power | ||||||
| Base | 125 W | 65 W | ||||
| Max Turbo | 250 W | 159 W | 159 W | 121 W | ||
The 270K Plus is effectively a $300 flagship killer.
The 270K Plus is effectively a $300 flagship killer. It retains the same 24-core, 24-thread configuration with 8 P-cores and 16 E-cores as the Core Ultra 9 285K. However, instead of asking $610, Intel has brought this configuration down into the Ultra 7 tier at a far more aggressive $300 price point. Availability has been limited so far, but in markets like Australia it can be found for around $530 AUD, roughly half the price of the 285K, representing a 46% discount.
As noted earlier, the original Arrow Lake processors suffered from noticeable gaming regressions due to higher latency between tiles. To address this, Intel significantly increased the die to die interconnect frequency, raising it from 2.1GHz to 3.0GHz. This 900MHz increase helps reduce inter tile latency, providing a much needed improvement for gaming workloads.
In our testing, the 270K Plus is able to deliver performance similar to the 14900K while using a fraction of the power. That is not an extraordinary result given how long it took to get here, but it is a much stronger showing for the Core Ultra 200 series.
Intel still has significant ground to cover, and if current reports are accurate, the company is already preparing a more aggressive response.
Nova Lake is Intel's next attempt to reclaim performance leadership.
The next generation, codenamed Nova Lake and expected to arrive as the Core Ultra 400 series, is shaping up to be one of the most substantial architectural shifts in recent years. It is expected to launch later in 2026 and is being positioned as a serious attempt to reclaim the performance lead in both multi-core productivity and gaming.
Nova Lake is rumored to introduce new Coyote Cove performance cores and Arctic Wolf efficient cores, both expected to deliver meaningful IPC gains over Arrow Lake. To reach significantly higher core counts, reports suggest a flagship configuration featuring 16 P-cores, 32 E-cores, and an additional 4 low power E-cores. This would be enabled by a "dual compute tile" design, effectively combining two CPU dies within a single package alongside the SoC and graphics tiles.
There are also indications that Intel is working to counter AMD's advantage in gaming with its X3D processors by introducing bLLC (Big Last-Level Cache), a larger last level cache design. High-end models are rumored to feature up to 144MB of additional L3 cache, while dual tile variants could reach as high as 288MB.
All of this sounds promising, and as expected, in typical Intel fashion it will arrive with a new platform. As we predicted when the Core Ultra 200 series launched, the LGA 1851 socket appears set to have a short lifespan. Nova Lake will force users to upgrade to an entirely new motherboard platform featuring the LGA 1954 socket, packing more pins to handle more bandwidth and power. The key question is whether this platform will be another two-generation offering, or will Intel finally get serious.
It seems as though Intel's business model has long relied on regular chipset refreshes for its motherboard partners. The company has not shown a willingness to support a single consumer socket for four to five years, and there is little to suggest LGA 1954 will be different. Still, there is some hope that this could mark a change in direction. For now, we wait and see.
That wraps up our revisit of Intel's decade of flagship processors. If you enjoyed all the testing we put into this feature, please share it, subscribe to our newsletter to receive news of future articles like this one, and check out our TechSpot Elite subscription option for ad-free browsing and additional perks.