Intel Core Ultra 5 245K, 235, 225: How to Choose? Processor Performance Unboxing

First model in the Intel Core Ultra 200S processor family to succeed the CPU overclocking capability Intel Core Ultra 9 285K, Ultra 7 265K, Ultra 5 245KF After that, the cheaper but non-overclockable Intel Core Ultra 200 Non-K series models will make their official debut at CES 2025 and will be available for purchase around mid-January of the same year. This time, we will actually test the Non-K series models that people are likely to buy, such as Intel Core Ultra 5 235 and Intel Core Ultra 5 225, for your reference, as well as the similarly priced Intel Core Ultra 5 245K. This time, we will test the Non-K series Intel Core Ultra 5 235 and Intel Core Ultra 5 225 for your reference, as well as the Intel Core Ultra 5 245K, which is in the same price range, so that you can make reference to how to choose the Intel Core Ultra 5 processor.

Intel Core Ultra 5 Base Specification 245K, 235, 225

This time, I would like to test the Intel Core Ultra 200 Non-K series models, in fact, the main reason is to deal with the NVIDIA RTX 50 series graphics cards that will be on sale in the coming year, RTX 5090 and 5080, "I heard" there may be a restriction on the sale of the whole machine assembly, and the sales channels will be open for the early bird project is to restrict the ICU platform, what will happen in the future the author can't predict the future. I can't predict what will happen in the future, but I think many people will want to assemble new machines in the coming years, and maybe some people will want to know how much performance difference there is between Intel Core Ultra 5 245K, Intel Core Ultra 5 235, and Intel Core Ultra 5 225, so I wrote this unboxing article.

Intel Core Ultra 200 Non-K series models were actually announced at CES 2025, but everyone was looking at NVIDIA and AMD's new products and no one cared about Intel, anyway, the first release of the Intel Core Ultra 9 285, Ultra 7 265, Ultra 5 245, and other 65W However, the only Non-K series available in Taiwan are Intel Core Ultra 5 235 and Intel Core Ultra 5 225(F), but nothing else for now.

Now all the common models of Intel Core Ultra 200S Series SKUs have been released, and the author would like to focus on the next point, most of the people who would buy the Non-K series are trying to save some budget, after all, not everyone wants to overclock when assembling a price-performance gaming machine or a desktop computer, so according to the logic of this thinking, the models to buy should be Intel Core Ultra 5 245, 235, 225(F).

The Intel Core Ultra 5 245 is not available now, and the Intel Core Ultra 5 235 is 14 cores (6 P-Core+8 E-Core) and 14 threads (14 threads), with a maximum single-core frequency of 5 GHz and 24 MB of L3 cache, which basically makes it the 245 with a 0.1 GHz cut in frequency and a 1 GPU Gores cut in the graphics card cores. It can basically be seen as a 245 with the frequency cut by 0.1 GHz and the built-in graphics cores cut by 1 GPU Gores.

The cheaper Intel Core Ultra 5 225 has 10 cores (6 P-Core+4 E-Core) and 10 threads (10 threads), with a maximum single-core frequency of 4.9 GHz and an L3 cache of 20 MB. Compared to the 235, the E-Core has been removed by four, and the frequency has been slashed by 0.1 GHz, and then the built-in graphics have been added. The E-Core has been removed from the 235, the frequency has been cut by 0.1 GHz, and the built-in graphics card core has been cut to 2 GPU Gores, so the rest of the specs are as follows.

∆ Intel Core Ultra 200S Series SKU(Non-K) series without processor overclocking.

∆ Detailed specifications are provided for your reference.

The ∆65W SKU comes with a factory fan specification, the Intel Core Ultra 9 is more advanced, but I'm sure we should all upgrade to a better looking cooler on our own.

 

Intel Core Ultra 5 245K, 235, 225 Calculation, renderingProfessional Performance Measurement

This time, we are testing three Intel Core Ultra 5 series processors, including the Intel Core Ultra 5 245K, 235, 225, etc. We are going to test them with the B860 chipset, which is more likely to be the mainboard that fits the budget constraints, and the graphics card is going to be used with the RTX 4060 Ti Founders Edition. The graphics card was paired with an RTX 4060 Ti Founders Edition for the test to make the overall configuration more in line with the logic of low-budget orders. After all, who would use a processor within the $10,000 range with a Z890 or a 4090 / 5090?

Memory is used Kingston FURY Renegade DDR5 RGB Limited Edition 48GB (2x24GB) Tested with the built-in XMP Profile 3 DDR5-6400 CL32-39-39-80 1.4V. Since I don't have any 6000 MT/s memory on hand, I used the existing memory to simulate the test by applying the built-in lower frequency profile parameters.

For the rest of the platform settings, except for turning on the XMP Profile and manually setting the fans to run at full speed, all other settings remain default, and the motherboard Power Delivery Profile stays in the default Intel Default Mode.

Intel Core Ultra 5 245K, 235, 225 Testing Platform

Processor:Intel Core Ultra 5 245K, 235, 225 (Power Delivery Profile_Intel Default Mode)
Radiator:FSP MP7 Black(full speed)
Motherboard: ASRock B860 Steel Legend WiFi (BIOS version: 1.21)
Memory:Kingston FURY Renegade DDR5 RGB Limited Edition 48GB (2x24GB) / XMP Profile 3 DDR5-6400 CL32-39-39-80 1.4V
Graphics: NVIDIA GeForce RTX 4060 Ti Founders Edition 8GB
Operating System: Windows 11 Professional 24H2
System Dish:Kingston NV3 2TB
Game Dish:XPG GAMMIX S70 PRO PCIe Gen4 x4 M.2 SSD 4TB
Power supply:MONTECH TITAN PLA 1000W
Case: STREACOM BC1 Benchtable V2
Graphics driver: GeForce Game Ready 566.36

 

firstly CPU-Z Viewing the hardware information of this test platform, Intel Core Ultra 5 245K processor has 14(6P+8E) cores and 14 threads, Intel Core Ultra 5 235 processor has 14(6P+8E) cores and 14 threads, Intel Core Ultra 5 225 processor has 10(6P+4E) cores and 10 threads, series code name is Arrow Lake using TSMC processor, motherboard uses ASRock B860 Steel Legend WiFi supporting PCI-E 5 0 lanes, BIOS is 121. Intel Core Ultra 5 225 processor with 14(6P+8E) cores and 14 threads, Intel Core Ultra 5 225 processor with 10(6P+4E) cores and 10 threads, series code name Arrow Lake using TSMC process, motherboard using ASRock B860 Steel Legend WiFi with PCI-E 5.0 lanes support, BIOS version 1.21, memory using Kingston FURY Renegade DDR5 RGB Limited Edition 8000MT/s dual-channel capacity. Edition 8000MT/s dual-channel capacity totaling 48GB and operating at 1.4V DDR5-6400 CL32-39-39-80 with XMP Profile 3.

At the same time, we also ran CPU-Z built-in test Version 17.01.64, and the results are for your reference. 245K is about 6% higher than 235, and 235 is about 42.5 % higher than 225, based on the results of multiple runs.

∆ Intel Core Ultra 5 245K_CPU-Z information at a glance and Version 17.01.64 built-in test score results.

∆ Intel Core Ultra 5 235_CPU-Z information at a glance and Version 17.01.64 built-in test score results.

∆ Intel Core Ultra 5 225_CPU-Z information at a glance and Version 17.01.64 built-in test score results.

 

AIDA64 Cache & Memory Benchmark It is mainly used to test the performance of processor caches and RAM. The read/write/copy bandwidth performance score represents the rate of transfer between CPU and memory, which means the efficiency of data throughput (the higher the score, the better), and the time delay score of accessing data indicates the responsiveness of the memory system.

This use Kingston FURY Renegade DDR5 RGB Limited Edition 48GB (2x24GB) Dual-channel memory, open this memory SPD built-in XMP Profile 3 DDR5-6400 CL32-39-39-80 1.4V to test, on the B860 motherboard, read speed, write speed, copy speed, latency results are as follows for reference.

Just what does it say on the CPU Type side? It's all Intel Core Ultra “3”, AIDA64 can really mix it up a bit.

∆ Intel Core Ultra 5 245K_AIDA64 Cache and Memory Test.

∆ Intel Core Ultra 5 235_AIDA64 Cache and Memory Test.

∆ Intel Core Ultra 5 225_AIDA64 Cache and Memory Test.

 

new version Cinebench 2024 MAXON ONE is Maxon's benchmark software for MAXON ONE software. MAXON ONE includes Cinema 4D, Red Giant Complete, Redshift, ZBrush, Universe, and Forger, and is a powerful tool for creating animation effects, motion design, motion graphics, film-grade portraits, and game art scenes. Each of these software programs fits together to provide a complete visual solution for creators.

Cinebench 2024 uses Cinema 4D's default Redshift rendering engine to test GPU and CPU performance. If multiple graphics cards are used to run Cinebench 2024 tests, the software utilizes multiple cards for the rendering tests at the same time, and Cinebench 2024 performs multiple rendering tests with six times the amount of computation when compared to Cinebench R23. Compared to Cinebench R23, the Cinebench 2024 multi-threaded rendering tests show a six-fold increase in the amount of computation for scenes. This reflects improvements in CPU performance and the higher hardware requirements that multimedia workers must now meet.

In this test the Intel Core Ultra 5 245K received Multi Core 1447 pts and Single Core 132 pts, the Intel Core Ultra 5 235 received Multi Core 1150 pts and Single Core 128 pts, and the Intel Core Ultra 5 225 received Multi Core 887 pts and Single Core 125 pts. The Intel Core Ultra 5 235 gets Multi Core 1150 pts and Single Core 128 pts, and the Intel Core Ultra 5 225 gets Multi Core 887 pts and Single Core 125 pts.

The difference in single-core results is not that big, but when looking at Multi Core, the 245K outperforms the 225 by about 63.1 %, while the 235 outperforms the 225 by about 29.6 %.

∆ Intel Core Ultra 5 245K_Cinebench 2024.

∆ Intel Core Ultra 5 235_Cinebench 2024.

∆ Intel Core Ultra 5 225_Cinebench 2024.

 

CrystalMark Retro 1.0.2 It is a comprehensive benchmark software that measures CPU, HDD, 2D graphics (GDI) and 3D graphics (OpenGL) performance. Developed by hiyohiyo and koinec, the authors of CrystalDiskMark and CrystalDiskInfo, the results are individual scores and not in any common units.

∆ Intel Core Ultra 5 245K_CrystalMark Retro 1.0.2.

∆ Intel Core Ultra 5 235_CrystalMark Retro 1.0.2.

∆ Intel Core Ultra 5 225_CrystalMark Retro 1.0.2.

 

PCMark 10 It also simulates test scenarios to determine the overall performance of the computer. The common basic functions include application launching, web browsing, and video conferencing tests, while productivity simulates document and spreadsheet writing, and the last item, video content creation, includes professional tests such as photo editing, video editing, and rendering.

∆ Intel Core Ultra 5 245K_PCMark 10 test.

∆ Intel Core Ultra 5 235_PCMark 10 test.

∆ Intel Core Ultra 5 225_PCMark 10 test.

 

3DMark score test

Next up, the 3DMark series, one of the most benchmarks for gaming scores, was used to compare the theoretical results of the same graphics platform with different processors through a series of tests with different picture quality and different GPU APIs.

3DMark CPU Profile This test will measure the performance of MAX, 16, 8, 4, 2 and 1 threads respectively, while the performance of 16 threads and above is more for 3D rendering or audio/video professional work, most of the mainstream DirectX 12 games can refer to the score of 8 threads, while the score of 4 and 2 threads are related to the old games developed with DirectX 9.

∆ Intel Core Ultra 5 245K_3DMark CPU Profile.

∆ Intel Core Ultra 5 235_3DMark CPU Profile.

∆ Intel Core Ultra 5 225_3DMark CPU Profile.

 

3DMark Fire Strike Includes two tests using the DirectX 11 GPU API, a physical test and a combined CPU and GPU test, each testing theoretical game performance at 1080p.

∆ Intel Core Ultra 5 245K_DirectX 11 Fire Strike.

∆ Intel Core Ultra 5 235_DirectX 11 Fire Strike.

∆ Intel Core Ultra 5 225_DirectX 11 Fire Strike.

 

3DMark Time Spy The DirectX 12 GPU API is also benchmarked for gaming at 2K 1440p.

∆ Intel Core Ultra 5 245K_DirectX 12 Time Spy.

∆ Intel Core Ultra 5 235_DirectX 12 Time Spy.

∆ Intel Core Ultra 5 225_DirectX 12 Time Spy.

 

UL Procyon video editing and picture editing test

UL Procyon Two benchmarks are used for testing, Video Editing Benchmark for video editing and Photo Editing Benchmark for photo editing, both of which use Adobe software for benchmarking.

Video Editing Benchmark The video editing test is conducted using Adobe Premiere Pro on the computer. For editing media workers, Adobe Pr has never had a standardized test reference, and in this benchmark test, two video files will be imported, edited, adjusted, and set up special effects, and then output in H.264 (Youtube 1080P) / H.265 (4K). Get the total score and turn on graphics card acceleration for the test.

In the video editing test, although the scores of the three processors vary from high to low depending on the model specification, the difference in scores is basically less than 1%.

∆ Intel Core Ultra 5 245K_UL Procyon Video Editing Benchmark video editing score.

∆ Intel Core Ultra 5 235_UL Procyon Video Editing Benchmark video editing score.

∆ Intel Core Ultra 5 225_UL Procyon Video Editing Benchmark video editing score.

 

Photo Editing Benchmark The first item of the photo editing test will import the Digital Negative into Adobe Lightroom Classic and then crop, stretch and modify the DNG image to test, while the second item will use Adobe Photoshop to apply multiple layers of editing effects and then export them to derive a photo retouching score / batch processing score. Batch Processing Score.

In the photo editing test, the 245K outperforms the 225 by about 5 %, while the 235 outperforms the 225 by about 3.8 %, so the difference in specifications is at least as significant as that in video editing.

∆ Intel Core Ultra 5 245K_UL Procyon Photo Editing Benchmark Photo Editing Score.

∆ Intel Core Ultra 5 235_UL Procyon Photo Editing Benchmark Photo Editing Score.

∆ Intel Core Ultra 5 225_UL Procyon Photo Editing Benchmark Photo Editing Score.

 

Intel Core Ultra 5 245K, 235, 225 threeFPS type gaming performance test

Three iconic games have been selected for testing in the FPS gaming category.Battle Attack 2_Overwatch 2, _Rainbow Six: Operation Siege_Tom Clancy's Rainbow Six Siege, and theApex LegendsThe APEX is set to the "highest" setting in the game, and the APEX is set to the maximum frame rate of 144 fps by default, so manually input the code to the maximum of 300 fps.

With a standardized resolution of Full HD 1920×1080 to allow for side-by-side comparisons of processor performance under stress, average (AVG) fps and 1% Low fps were collected through NVIDIA GeForce's free FrameView software.

《Rainbow Six: Operation SiegeTom Clancy's Rainbow Six Siege: NVIDIA Reflex Enabled + Enhanced, Overall Quality_Super High
《Apex Legends":DirectX 12, NVIDIA Reflex Enablement + Enhancements, TEXTURE STREAMING BUDGET_LIMITED (8GB VRAM)
《Battle Attack 2_Overwatch 2: NVIDIA Reflex Open + BOOTS, NVIDIA DLSS Super Resolution

Of the three FPS-type e-sports games, except for Tom Clancy's Rainbow Six Siege, which has a built-in performance test, the other two games are all about collecting frames at the shooting range.

• Average fps: Record the number of frames generated during a test play session and divide by the total recording time to get the average value.
• 1% Low fpsThe 1% part of the frame value that represents the longest frame generation time (Frametime) during the test is calculated as the average value, the closer the value is to the measured average fps, it means that the game does not have a significant frame drop problem, in other words 1% Low fps is the fluctuation of the low frame rate.

∆ Three FPS-type gaming games tested: highest texture effects, FHD 1080×1920 resolution.

∆ Comparison of the total average fps of the three FPS-type gaming tests.

 

Intel Core Ultra 5 245K, 235, 225 十AAA Masterpiece Gaming Performance Tests

To test the AAA genre, which emphasizes on the experience of graphic performance and storytelling, the author chose to play "The Battle of the Birds", which is the first game in the AAA genre.Bloodshed 2_ Red Dead Redemption 2, _Ghosts of Tsushima Warriors_Ghost of Tsushima》、《Electric Renegade 2077_Cyberpunk 2077, TheBrothers of the Four Seas: The Definitive Edition_Mafia: Definitive Edition, theHorizon: The Western Sanctuary_ Horizon Forbidden West, _The Witcher 3: Wild Hunt_The Witcher 3: Wild Hunt, TheKurokami Dialog: Goku_Black Myth: Wukong, TheWatch Dogs: Freedom Corps_Watch Dogs: Legion, TheHowl of War 6_Far Cry 6, _Assassin's Creed: Illusions_Assassin's Creed Mirage" and ten other 3A masterpieces are tested in the game.

《Bloodshed 2_Red Dead Redemption 2: Quality preset (quality-focused), NVIDIA DLSS Quality, DirectX 12
《Ghosts of Tsushima Warriors_Ghost of Tsushima":Default_Very High, Boost_Method_DLSS, Boost_Quality, Framing_DLSS Framing
《Cyberpunk 2077":DLSS Super Resolution Default_Quality / Transforner Model, DLSS Frame Generation On, Material Quality_High, Light Tracing_Crazy
《Brothers of the Four Seas: The Definitive Edition_Mafia: Definitive Edition》:Total Quality Preset_Best
《Horizon: The Western Sanctuary_Horizon Forbidden West: DLSS quality, DLSS frame generation on, preset_very high
《The Witcher 3: Wild HuntThe Witcher 3: Wild Hunt: DirectX 12, Default Image_Light Following: Extreme, DLSS Quality, DLSS Frame Generation_On
《Kurokami Dialog: GokuBlack Myth: Wukong: DLSS Quality, Frame Generator_On, Panoramic Light Tracking_On/Ultra High, Quality Class_Video Class
《Watch Dogs: Freedom Corps_Watch Dogs: Legion: DirectX 12, Screen Quality_High, Light Tracing Reflections_Maximum, DLSS_Picture Mode
《Howl of War 6_Far Cry 6_: High Quality, DXR Light Tracing Reflections_on, DXR Light Tracing Shadows_on, FSR 1.0_Quality
《Assassin's Creed: IllusionsAssassin's Creed Mirage: World Detail_Very High, Upper Sample Type_NVIDIA DLSS

Considering the fact that most players pursue the most beautiful screen performance rather than the highest frame rate performance when playing 3A masterpiece games, the author also set the built-in "Highest/Higher" texture effect option for AAA type games, and chose "Highest" and DLSS 3-quality quality mode for the light tracking setting, to obtain a certain degree of frame rate enhancement while keeping the screen performance as much as possible, and the frame rate collection. Also uses Software Average Recording (AVG) and 1% Low fps.

• Average fps: Record the number of frames generated during a test play session and divide by the total recording time to get the average value.
• 1% Low fpsThe 1% part of the frame value that represents the longest frame generation time (Frametime) during the test is calculated as the average value, the closer the value is to the measured average fps, it means that the game does not have a significant frame drop problem, in other words 1% Low fps is the fluctuation of the low frame rate.

∆ Ten AAA games tested: highest texture effects, FHD 1080×1920 resolution.

∆ Comparison of the total average fps of the ten AAA games tested.

 

Intel Core Ultra 5 245K, 235, 225 Processor Temperature and Power Consumption Tests

The test platform was set up on a bare indoor test platform in a closed room at 20~22 °C. The ambient temperature in a normal room is difficult to control, so it is for reference only.

The author used Cinebench 2024 (Multi Core_one round), AIDA64 FPU, AIDA64 CPU, theKurokami Dialog: GokuBlack Myth: Wukong" and other scenarios were tested, and data was collected using HWiFO64 to collect and record the maximum temperature, and convert the information into graphs for players' reference.

1. AIDA64 FPU_30 Minute
2. AIDA64 CPU_30 Minute
3. Black Myth: Wukong 1920 x 1080(FHD)_Performance Test Tool 1 Round
4. Cinebench 2024 Multi Core_1 Round

• CPU temperature is measured by the diode in the slot to measure the internal (core) or external (case) temperature, but there is no way to know exactly whether it is the internal or external temperature that is measured, and most monitoring software uses this item to display the CPU temperature.
• CPU Package temperature is recorded by all Digital Temperature Sensors (DTS) in the package as the average of the maximum temperature over 256 milliseconds, and is the HWiNFO64 recommended CPU temperature observation value, which is also used as a benchmark for overheating and downclocking of the processor.

∆ Intel Core Ultra 5 245K_ thermal performance test and power consumption performance.

∆ Intel Core Ultra 5 235_ thermal performance test and power consumption performance.

∆ Intel Core Ultra 5 225_ thermal performance test and power consumption performance.

 

Conclusion

First of all, let's talk about these processors from the price point of view. Most of the people who will buy Intel Core Ultra 200 Non-K series models will consider them because they don't need to overclock and the price is cheaper than the K series, so they will consider them in order to pursue a more affordable setup list, however, the prices of the Intel Core Ultra 5 245K, 235, and 225 that the author tested are $9999, $9300, and $8750, respectively. Prices are $9,999, $9,300, and $8,750 respectively. The cheapest Intel Core Ultra 5 225F is $8,200 without the built-in graphics card, and the newest Intel Core Ultra 5 225 is $100 cheaper than the 9600X, which is the R5 9600X, compared to the contemporary AMD competitor. The Intel Core Ultra 5 225 is $100 cheaper than the 9600X in Cinebench 2024 Multi Core performance, and loses out to its price rival, while the Ultra 5 235 Multi Core performance is a little better but $450 more expensive.

In addition to contemporary competitors, we can also look back at the previous generation models. In fact, there is still the Ryzen 7 7700 MPK in the channel for consideration, its Cinebench 2024 score is basically the same as the R5 9600X, and its performance is also sandwiched between the Intel Core Ultra 5 235 and 225, but the R7 7700 MPK is even cheaper than the Intel Core Ultra 5 225F, about $ 1,200. But the price of the R7 7700 MPK is even cheaper than the Intel Core Ultra 5 225F by about $1,200. Considering the fact that AMD's product line has a lifecycle that can be continually reborn from the ground up, and that it can update the BIOS on the same board to upgrade the CPU in the next couple of years, as well as the price of the previous generation of AMD's 600-series chipset motherboards, etc., the answer is pretty obvious. Core Ultra 5 Non-K SKU is too expensive!

After looking at the competitor comparisons we go back to the Intel Core Ultra 5 245K, 235 and 225 processors on the B860 motherboard. We also look at the Cinebench 2024 rendering test as a comparison, and there is not a big difference in the single core results of the three processors, while when we look at the multi-core results, the 245K outperforms the 225 by about 63.1 % and the 235 outperforms the 225 by about 29.6 %. The 245K outperforms the 225 by about 63.1 %, while the 235 outperforms the 225 by about 29.6 %.

If you're building a computer for video editing, in the Adobe Premiere Pro test using UL Procyon and supporting graphics acceleration, the results of the three processors are basically less than 1% although the results vary from high to low depending on the model specification, so you can choose the cheapest Intel Core Ultra 5 225 (F) and invest some more budget in the graphics card. (F) and then invest some more budget on the graphics card, after all, it is your food equipment.

UL Procyon Photo Editing Benchmark photo editing test using Adobe Lightroom Classic and Adobe Photoshop, 245K than 225 up about 5 % results; 235 than 225 up about 3.8 % results, the specification difference brings at least more significant growth than the video editing. The difference in specifications is at least as significant as the increase in video editing, but it's a matter of opinion whether the increase is worth the extra thousand dollars.

The gaming performance part of the test this time with RTX 4060 Ti, before RTX 5060 and other graphics cards come out of this budget range of consumers at most to choose this model, if the author test of these three FPS gaming type games, Intel Core Ultra 5 225 can already bring 4060 Ti, 245K total average fps than 225 about 2.5%; total average 1% Low fps is 8.8%. The total average fps of the 245K is about 2.5% more than the 225; the total average 1% Low fps is 8.8%, and the ten 3A masterpiece (AAA type) games pursuing the quality of the graphics are basically no different in most of the items, and there is no difference only in the "The Best of the Best".Watch Dogs: Freedom Corps_Watch Dogs: Legion and theAssassin's Creed: IllusionsAssassin's Creed Mirage" has a big difference, I guess I will play these 3A masterpieces more pressure on the graphics card, so the 4060 Ti with Intel Core Ultra 5 225 is already enough.

Cooling temperature and power consumption of the three processors under the test at most 154W and already, spend a thousand dollars or so to choose a dual-tower six heat pipe air-cooled tower dissipation is already very enough, anyway, I do not want to use Intel's original cooler, sorry but the appearance really does not attract me.