AMD Ryzen AI Max 390 vs Intel Core Ultra 5 228V vs Intel Core Ultra 7 256V
AMD Ryzen AI Max 390
► remove from comparison
The AMD Ryzen AI Max 390 is a powerful Strix Halo family processor that debuted in January 2025. The APU comes equipped with 12 Zen 5 CPU cores running at up to 5.0 GHz, the 32 CU RDNA 3+ Radeon 8050S graphics adapter and the 50 TOPS XDNA 2 neural engine. Other noteworthy features include PCIe 4, USB 4 and LPDDR5x-8000 RAM support and a whole lot of L3 cache.
Architecture and Features
Unlike it is with Strix Point, Strix Halo parts are powered by Zen 5 cores - no Zen 5c here. It's not clear if this is the desktop Zen 5 implementation with full AVX512 throughput or the mobile one. According to AMD, Zen 5 delivers a 16% IPC improvement over Zen 4 thanks to branch prediction improvements and other refinements.
Elsewhere, the AI Max chip supports RAM as fast as LPDDR5x-8000, and is natively compatible with USB 4 (and therefore Thunderbolt). It has PCIe 4.0 support for a throughput of 1.9 GB/s per lane, just like its 8000 series predecessors had. The integrated XDNA 2 NPU delivers up to 50 INT8 TOPS for accelerating various AI workloads.
Performance
Its CPU performance should be about 10% higher than that of the Ryzen AI 9 HX 370 and HX 375 due to the AI Max chip's higher TDP power target and lack of cut-down Zen 5c cores.
Graphics
The Radeon 8050S sports 32 RDNA 3+ architecture CUs (2048 unified shaders) that could see it competing with lower mid-range desktop graphics cards like the Radeon RX 7700 XT. This GPU will undoubtedly run any game at 1080p on Ultra, however, the ultimate question is whether or not the cooling solution of the laptop will be powerful enough to let the iGPU shine.
Naturally, the Radeon is capable of driving four SUHD 4320p60 monitors. It can also efficiently encode and decode the most popular video codecs including AVC, HEVC, VP9 and AV1. The latest addition to that list, the VVC codec, is not supported unlike it is with Intel Lunar Lake chips.
Power consumption
The AI Max 390 can eat up to 120 W depending on the system and its TDP power targets, with 45 W named as the minimum TDP.
The 4 nm TSMC process that the CPU cores are built with make for decent, as of Jan 2025, energy efficiency.
Intel Core Ultra 5 228V
► remove from comparison
The Intel Core Ultra 5 228V is one of the 2 slowest Lunar Lake family processors as of Nov 2024. This is an SoC for use in tablets and laptops of the slimmer kind that was unveiled in Autumn 2024. It features 4 new Skymont E-cores and 4 new Lion Cove P-cores running at up to 3.5 GHz and 4.5 GHz respectively, along with the new Arc 130V iGPU and 32 GB of on-package LPDDR5x-8533 RAM. A new 40 TOPS neural engine, Thunderbolt 4 and PCIe 5 SSD support are included as well.
The only difference between this chip and the Ultra 5 226V is the amount of on-package, non-replaceable RAM: 32 GB vs 16 GB respectively.
Architecture and Features
Lunar Lake is built using the Foveros technology (stacking several dies on top of each other and next to each other), just like Meteor Lake was. The new chips make use of the enormous BGA2833 socket interface. Of the 8 cores, not a single one is Hyper-Threading-enabled which is the opposite of what AMD currently does with its Zen 5/5c chips.
Intel claims Lion Cove cores bring a 14% IPC improvement over Redwood Cove. For Skymont and Crestmont, the difference is much higher at 68%. Several tweaks and improvements are present here, such as the Low Latency Fabric that is supposed to make small data transfers between cores/caches a lot faster. The 228V's level 3 cache is pretty small as compared to more expensive Ultra 7 2xxV and Ultra 9 2xxV chips at 8 MB vs 12 MB. Elsewhere, the 228V has 4 PCIe 5 and 4 PCIe 4 lanes for connecting various kinds of devices, including NVMe SSDs at up to 15.75 GB/s. Thunderbolt 4 support is onboard by default, as is support for CNVi WiFi 7 + BT 5.4 cards from Intel. The 40 TOPS "AI Boost" neural engine is present along with technologies such as Threat Detection to make AI-enabled applications such as the Windows Defender more powerful.
Intel is predicted to get short of on-package RAM in subsequent CPU generations.
Performance
Our in-house testing of an Acer laptop powered by the 226V which is essentially the same chip revealed the latter to be about as fast as the Ryzen 7 7730U and the Ryzen 7 7735U in multi-threaded tasks which is OK but not mind-blowing for late 2024. It is worth noting that the laptop in question does not use the Intel-recommended 17 W long-term power target; instead, the processor gets 30 W to eat which affects its benchmark scores significantly.
Generally speaking, the 226V, 228V, 236V and 238V are slower than the 256V, 258V, 266V, 268V, 288V due to the difference in their last-level cache size as well as clock speeds. However, the difference in performance between the slowest Lunar Lake chip, the 226V, and the fastest chip, the 288V is fairly small at around 10% to 15%. Exact figures depend on the TDP values of the laptops being pitted against each other more than on anything else.
Graphics
The Arc Graphics 130V has 7 Xe² architecture "cores" running at up to 1,850 MHz, as well as 7 ray tracing units at its disposal. This is a direct successor to the Arc 7 iGPU; it is DirectX 12 Ultimate-enabled and able to HW-decode a long list of video codecs such as h.266 VVC, h.265 HEVC, h.264 AVC, AV1 and VP9. Three SUHD 4320p monitors can be used simultaneously with this iGPU.
Its performance can be vastly different depending on which benchmarks and games one runs. Oftentimes, it isn't much faster than its direct predecessor, the Arc 7. It is clear the 130V isn't fast enough to run most triple-A 2024 games at 1080p. However, it's still more than good enough for an occasional gaming session.
- Helldivers 2: 25 fps (Med, 1080p). The Arc 7 and the Radeon 880M deliver similar results.
- Ready or Not: 40 fps (Med, 1080p). The GeForce MX350 delivers a similar result.
Power consumption
This 2nd generation Core Ultra processor is supposed to consume 17 W when under long-term workloads. The Intel-recommended short-term power limit for the chip sits at 37 W.
Intel Core Ultra 7 256V
► remove from comparison
The Intel Core Ultra 5 256V is an upper mid-range Lunar Lake family processor. This is an SoC for use in tablets and laptops of the slimmer kind that was unveiled in Autumn 2024. It sports 4 new Skymont E-cores and 4 new Lion Cove P-cores running at up to 3.7 GHz and 4.8 GHz respectively, along with the new Arc 140V iGPU and 16 GB of on-package LPDDR5x-8533 RAM. A new 47 TOPS neural engine, Thunderbolt 4 and PCIe 5 SSD support are included as well.
The only difference between this chip and the Ultra 7 258V is the amount of on-package, non-replaceable RAM: 16 GB vs 32 GB respectively.
Architecture and Features
Lunar Lake is built using the Foveros technology (stacking several dies on top of each other and next to each other), just like Meteor Lake was. The new chips make use of the enormous BGA2833 socket interface. Of the 8 cores, not a single one is Hyper-Threading-enabled which is the opposite of what AMD currently does with its Zen 5/5c chips.
Intel claims Lion Cove cores bring a 14% IPC improvement over Redwood Cove. For Skymont and Crestmont, the difference is much higher at 68%. Several tweaks and improvements are present here, such as the Low Latency Fabric that is supposed to make small data transfers between cores/caches a lot faster. The 256V has a very healthy 12 MB of level 3 cache; elsewhere, it has 4 PCIe 5 and 4 PCIe 4 lanes for connecting various kinds of devices, including NVMe SSDs at up to 15.75 GB/s. Thunderbolt 4 support is onboard by default, as is support for CNVi WiFi 7 + BT 5.4 cards from Intel. The 47 TOPS "AI Boost" neural engine is present along with technologies such as Threat Detection to make AI-enabled applications such as the Windows Defender more powerful.
Intel is predicted to get short of on-package RAM in subsequent CPU generations.
Performance
Based on the time we've had with 2 systems powered by the 256V, the chip's multi-thread performance is as good as that of the Core Ultra 7 165U and the Core i7-1360P. Please note that both systems had a higher long-term TDP power target than the Intel-recommended value: 30 W and 32 W vs 17 W.
Generally speaking, the 288V, 268V, 266V, 258V, 256V are faster than the 238V, 236V, 228V and 226V due to the difference in their last-level cache size (8 MB vs 12 MB) as well as clock speeds. However, the difference in performance between the slowest Lunar Lake chip, the 226V, and the fastest chip, the 288V is fairly small at around 10% to 15%. It depends on the TDP figures of the laptops being pitted against each other more than on anything else.
Graphics
The Arc Graphics 140V is here to replace the Arc 8 iGPU. Its 8 Xe² architecture "cores" run at up to 1,950 MHz and it also has 8 ray tracing units at its disposal. The adapter is DirectX 12 Ultimate-enabled and able to HW-decode a long list of popular video codecs such as h.266 VVC, h.265 HEVC, h.264 AVC, AV1 and VP9. Three SUHD 4320p monitors can be used simultaneously with this iGPU.
All 2023 and 2024 games are playable at 1080p on low graphics settings or higher with this iGPU. We got well over 30 fps in Ghost of Tsushima and almost 40 fps in Baldur's Gate 3. This means the Radeon 780M gets left behind while the Radeon 890M reigns supreme.
Power consumption
This 2nd generation Core Ultra processor is supposed to consume 17 W when under long-term workloads. The Intel-recommended short-term power limit for the chip sits at 37 W.
Model | AMD Ryzen AI Max 390 | Intel Core Ultra 5 228V | Intel Core Ultra 7 256V | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Codename | Strix Halo | Lunar Lake | Lunar Lake | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Series | AMD Strix Halo/Point (Zen 5/5c, Ryzen AI 300) | Intel Lunar Lake | Intel Lunar Lake | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Series: Lunar Lake Lunar Lake |
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Clock | 3200 - 5000 MHz | 2100 - 4500 MHz | 2200 - 4800 MHz | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
L2 Cache | 12 MB | 14 MB | 14 MB | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
L3 Cache | 64 MB | 8 MB | 12 MB | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Cores / Threads | 12 / 24 12 x 5.0 GHz AMD Zen 5 | 8 / 8 4 x 4.5 GHz Intel Lion Cove P-core 4 x 3.5 GHz Intel Skymont E-core | 8 / 8 4 x 4.8 GHz Intel Lion Cove P-core 4 x 3.7 GHz Intel Skymont E-core | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TDP | 55 Watt | 17 Watt | 17 Watt | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Technology | 4 nm | 3 nm | 3 nm | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Socket | FP11 | BGA2833 | BGA2833 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Features | LPDDR5x-8000 RAM, PCIe 4, USB 4, XDNA 2 NPU (50 TOPS), SMT, AES, AVX, AVX2, AVX512, FMA3, MMX (+), SHA, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, SSE4A | LPDDR5x-8533 RAM, PCIe 5 + PCIe 4, USB 4, AI Boost NPU (40 TOPS), Thread Director, PSE, Threat Detection, MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, AES, AVX, AVX2, AVX-VNNi, FMA3, SHA | LPDDR5x-8533 RAM, PCIe 5 + PCIe 4, USB 4, AI Boost NPU (47 TOPS), Thread Director, PSE, Threat Detection, MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, AES, AVX, AVX2, AVX-VNNi, FMA3, SHA | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
iGPU | AMD Radeon RX 8050S ( - 2800 MHz) | Intel Arc Graphics 130V ( - 1850 MHz) | Intel Arc Graphics 140V ( - 1950 MHz) | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Architecture | x86 | x86 | x86 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Announced | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Manufacturer | www.amd.com | ark.intel.com | ark.intel.com | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
L1 Cache | 1.4 MB | 1.4 MB | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TDP Turbo PL2 | 37 Watt | 37 Watt | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
max. Temp. | 100 °C | 100 °C |
Benchmarks
Average Benchmarks AMD Ryzen AI Max 390 → 100% n=3
Average Benchmarks Intel Core Ultra 5 228V → 85% n=3
Average Benchmarks Intel Core Ultra 7 256V → 91% n=3

* Smaller numbers mean a higher performance
1 This benchmark is not used for the average calculation