CPUs / Compare

AMD Ryzen AI Max+ 395 vs Intel Core Ultra X9 388H

A

Product A: AMD

Ryzen AI Max+ 395

43 of 74 known

Change AMD Ryzen AI Max+ 395

Compare the Intel Core Ultra X9 388H with

Search CPUs

B

Product B: Intel

Core Ultra X9 388H

47 of 74 known

Change Intel Core Ultra X9 388H

Compare the AMD Ryzen AI Max+ 395 with

Search CPUs

All 74 fields

Every value links to where we found it.

Only differences

AAMD Ryzen AI Max+ 395

BIntel Core Ultra X9 388H

General

6 of 13 differ

Series

Ryzen AI Max 300 Series1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Series row: Ryzen AI Max 300 Series.

Core Ultra Series 34Press releaseintel.com/content/www/us/en/newsroom/news/artificial-intelligence/ces-2026-intel-core-ultra-series-3-debut-first-built-on-intel-18a.htmlRead 15 October 2026Intel calls the family Intel Core Ultra Series 3. ARK's Product Collection reads Intel Core Ultra Series 3 processors.

Model numbersiMaker identifiers for this exact model and its listed variants. Excludes sibling models and other generations.

100-0000010991Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Product ID Tray 100-000001099. AMD lists no boxed product.

QQ807293B9B40E3, QQ807293B9B45852Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/ordering.htmlRead 15 October 2026Ordering codes QQ807293B9B40E3 (spec code Q9EK) and QQ807293B9B4585 (spec code SA4QW), both Tray, stepping B0, package FCBGA20X. ARK lists no boxed product.

GTIN / EAN

Not available

Not available

AnnouncediThe maker's first announcement of this model. This is separate from first retail availability.

6 January 20252Press releaseir.amd.com/news-events/press-releases/detail/1232/amd-announces-expanded-consumer-and-commercial-ai-pc-portfolioRead 15 October 2026AMD release dated 6 January 2025 (ahead of CES) announces the Ryzen AI Max Series and lists the Ryzen AI Max+ 395 with 16C/32T, up to 5.1 GHz boost, 3.0 GHz base, 80 MB total cache, Radeon 8060S, 45-120W cTDP, 50 NPU TOPS and 40 graphics cores. It says systems are expected to be available starting in Q1 2025 and gives no price.

5 January 20264Press releaseintel.com/content/www/us/en/newsroom/news/artificial-intelligence/ces-2026-intel-core-ultra-series-3-debut-first-built-on-intel-18a.htmlRead 15 October 2026Intel release dated 5 January 2026 (CES, Las Vegas) introduces the Intel Core Ultra Series 3 processors. It does not name the 388H, and says the top SKUs have up to 16 CPU cores, 12 Xe-cores and 50 NPU TOPS.

Status

Available1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Product is listed and current on AMD's page, with Regional Availability Global.

Available1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Marketing Status: Launched. Launch Date: Q1'26. Product Collection: Intel Core Ultra Series 3 processors. Vertical Segment: Mobile.

PredecessoriThe model this one directly replaced, in the same line, tier and size, as the maker or a trusted review names it.

Not available

Not available

SuccessoriThe model that directly replaced this one, in the same line, tier and size. A revision such as a Slim or OLED model does not count when a next generation exists.

Not available

Not available

Launch price (US)iThe maker's US launch MSRP for the stated configuration, excluding tax. Current store prices and later discounts do not qualify.

Not available

Not available

Launch price (EU)iThe maker's euro launch price for the stated configuration, including VAT as published. Current store prices and later discounts do not qualify.

Not available

Not available

MarketsiISO country or region codes where this exact model or variant is sold. A source page's language does not establish availability.

Not available

Not available

ColoursiThe maker's own colour names

Not available

Not available

SegmentiThe segment the chip maker sells it in. Intel calls laptop chips Mobile. AMD laptop chips that also list Desktops, for mini PCs, stay laptop.

Laptop1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Form Factor lists Laptops and Desktops. AMD laptop chips that also list Desktops stay laptop.

Laptop1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Vertical Segment row reads Mobile.

Instruction setiApple and Qualcomm chips are Arm64. Intel and AMD chips are x86-64.

x86-641Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Supported Extensions list x86-64.

x86-641Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Instruction Set: 64-bit, with Intel 64 listed as Yes. Intel's 64-bit processors are x86-64.

CodenameiThe chip maker's codename. Intel: the Code Name row without "Products formerly". AMD: the Former Codename row without the socket.

Strix Halo1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Former Codename row reads Strix Halo.

Panther Lake1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Code Name row reads Products formerly Panther Lake.

CPU process (as published)iVerbatim name of the process the CPU cores are made on, e.g. TSMC N3B or Intel 18A.

TSMC 4nm FinFET1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Processor Technology for CPU Cores: TSMC 4nm FinFET. The I/O die is also TSMC 4nm FinFET. Package Die Count is 3.

Intel 18A1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026CPU Lithography row: Intel 18A. The GPU tile lithography row reads TSMC N3E.

CPU process nodeiThe nm class of the named CPU-core process, as on laptops and desktop PCs: TSMC N3B is 3, Intel 7 is 10, Intel 4 is 7 and Intel 18A is 1.8. The exact name is in CPU process (as published). I/O and graphics dies may use other nodes.

4nm1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026TSMC 4nm FinFET is stored as 4 nm.

1.8nm1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Intel 18A is stored as 1.8 nm by the field rule. Intel prints no nm figure.

AVX-512iThe chip maker lists AVX-512 among the supported instruction extensions.

Yes1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Supported Extensions list includes AVX512.

No1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Instruction Set Extensions lists Intel SSE4.1, SSE4.2 and AVX2 only.

Cores and clocks

4 of 7 differ

Core configurationiOne row per core group the chip maker names. Intel names performance, efficient and low-power efficient cores. AMD names Zen 5 and Zen 5c cores. Qualcomm names prime and performance cores.

AAMD Ryzen AI Max+ 395

Count	Core	Base clock (GHz)	Max clock (GHz)
16	Zen 5	3	5.1

Source1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026AMD names no core roles. Processor Architecture reads 16x Zen 5 and # of CPU Cores is 16. Base Clock 3 GHz is AMD's average effective base frequency of all cores. Max. Boost Clock is the highest single-core clock, Up to 5.1 GHz.

BIntel Core Ultra X9 388H

Core type	Count	Core	Base clock (GHz)	Max clock (GHz)
Performance	4	Cougar Cove	2.1	5.1
Efficiency	8	Darkmont	1.6	3.8
Low-power efficiency	4	Darkmont	1.6	3.7

Source1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026# of Performance-cores 4 with base 2.1 GHz and Max Turbo 5.1 GHz. # of Efficient-cores 8 with base 1.6 GHz and Max Turbo 3.8 GHz. # of Low Power Efficient-cores 4 with base 1.6 GHz and Max Turbo 3.7 GHz. Total cores 16. ARK does not name the core architectures. Cougar Cove and Darkmont come from the Chips and Cheese and Igor's Lab Panther Lake architecture articles.5Press articleold.chipsandcheese.com/2025/10/14/panther-lakes-reveal-at-itt-2025Read 15 October 2026Chips and Cheese, 14 October 2025: the P-cores are codenamed Cougar Cove and the E-cores Darkmont. The largest Panther Lake configuration has 4 P-cores, 8 E-cores and 4 LP E-cores, with the 18 MB L3 shared by the P-cores and E-cores and the 12 Xe3 core GPU tile.6Press articleigorslab.de/en/evolution-or-revolution-intel-panther-lake-including-architecture-efficiency-and-software-integrationRead 15 October 2026Igor's Lab names the compute tile cores as four Cougar Cove-P, eight Darkmont-E and four Darkmont-LP-E cores. The LP E-core row stores Darkmont from this source.

CoresiPhysical cores, not threads.

16=CalculatedFrom Core configuration

16=CalculatedFrom Core configuration

Cores labelled performanceiCounts only cores the chip maker labels performance. Not available when the maker labels no roles, as on AMD chips. Qualcomm and Apple use other labels for their fastest cores.

Not available

4=CalculatedFrom Core configuration

Cores labelled efficiencyiCounts efficient and low-power efficient cores. Not available when the maker labels no roles.

Not available

12=CalculatedFrom Core configuration

ThreadsiHardware threads, including simultaneous multithreading where the chip has it.

321Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Multithreading (SMT) Yes, # of Threads 32.

161Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Total Threads 16. There are no multi-threaded cores in this layout: 16 cores and 16 threads.

Max boost clockiHighest clock the chip maker states for the processor, Intel Max Turbo Frequency or AMD Max. Boost Clock. Often reached on one or two cores only. It can be higher than every core group's max clock.

5.1GHz1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Max. Boost Clock: Up to 5.1 GHz.

5.1GHz1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Max Turbo Frequency 5.1 GHz, the same as the Performance-core Max Turbo Frequency and the Turbo Boost Max Technology 3.0 Frequency.

Unlocked for overclockingiThe core multiplier is unlocked, as on Intel K models and AMD chips listed as Unlocked for Overclocking. Memory profiles such as XMP or EXPO alone do not count.

Not available

Not available

Cache

2 of 3 differ

L2 cacheiTotal L2 cache of all cores.

16MB1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026L2 Cache 16 MB. AMD's launch table lists 80 MB total cache, which is L2 plus L3.

Not available

L3 cacheiTotal L3 cache, including 3D V-Cache. Intel Smart Cache is L3. L2 is not added. A total cache figure from Qualcomm or Apple is not L3.

64MB1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026L3 Cache 64 MB.

18MB1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 202618 MB Intel Smart Cache. The Total L2 Cache row is not on ARK.

AMD 3D V-CacheiExtra L3 cache stacked on a compute die. Yes when at least one compute die has it.

No1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026AMD lists 64 MB L3 and no stacked cache or 3D V-Cache technology.

No1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026AMD technology. ARK lists only the 18 MB Smart Cache and no stacked cache.

Power and thermals

3 of 4 differ

Base power (TDP)iIntel Processor Base Power or AMD Default TDP. A rating for cooler and board design, not measured power draw. A Default TDP printed as a range, such as 35-54 W, is not stored.

55W1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Default TDP 55W.

25W1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Processor Base Power 25 W.

Max turbo poweriHighest package power limit at stock settings. Intel Maximum Turbo Power, or AMD's package power tracking (PPT) limit. The source note says which. Only a source that names this exact chip counts. A database value applied to a whole class of chips does not.

Not available

80W1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Maximum Turbo Power 80 W with no footnote on ARK. Intel's launch SKU slide and the Quick Reference Guide print two values for this chip, 65 W and 80 W, which means the limit depends on the laptop maker's configuration. The ARK value is stored.3Maker spec sheet (PDF)cdrdv2-public.intel.com/871380/Intel%20Core%20Ultra%20Series%203%20Processors%20-%20Quick%20Reference%20Guide%20v1.pdfRead 15 October 2026Page 3 lists Maximum Turbo Power 65, 80 W with the Processor Base Power 25 W. The footnote says systems with a base power of 25 W and a turbo power of 64 W are supported, and systems with a base power of 45 W and a turbo power of 80 W. The ARK value of 80 W is stored.

Configurable power rangeiThe sustained power range a PC or laptop maker may set. AMD cTDP. Intel Minimum and Maximum Assured Power.

Minimum45Maximum120W1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026AMD Configurable TDP (cTDP) 45-120W.

Minimum15W1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Minimum Assured Power 15 W. ARK has no Maximum Assured Power row, so max_w is left out.

Max operating temperatureiHighest die temperature the chip maker allows, called Tjmax or Tjunction. The chip lowers its clocks to stay below it.

100°C1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Max. Operating Temperature (Tjmax) 100C.

100°C1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Max Operating Temperature 100 C, listed under Package Specifications.

Memory

5 of 6 differ

Memory supportiOfficial memory speeds per memory type, module type and module configuration, as the chip maker states them.

AAMD Ryzen AI Max+ 395

Memory type	Module type	Max speed (MT/s)
LPDDR5X	Soldered	8,000

Source1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026System Memory Type: 256-bit LPDDR5x. Max Memory Speed: LPDDR5x-8000. AMD states no module configuration. The memory is soldered to the board.

BIntel Core Ultra X9 388H

Memory type	Max speed (MT/s)
LPDDR5X	9,600

Source1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Memory Types row: Up to LPDDR5X 9600 MT/s. ARK lists no DDR5 and names no module type, rank or configuration, so those keys are left out.

Max memory speed (official)iHighest official speed in the memory support table, for any module type. Four modules or plain UDIMMs can run slower. Intel quotes some chips with CUDIMM speeds only.

8,000MT/s=CalculatedFrom Memory support

9,600MT/s=CalculatedFrom Memory support

Memory channelsiCounted in 64-bit channels, as Intel and AMD count them. A DDR5 module has two 32-bit subchannels and makers count it as one channel. A maker that counts narrower channels and states their width is converted by the total bus width, so Qualcomm's 8 channels of 16 bits are 2. A total bus width alone is divided by 64, so AMD's 256-bit interface is 4. The maker's own count goes in the source note.

44Press articletomshardware.com/pc-components/cpus/amds-beastly-strix-halo-ryzen-ai-max-debuts-with-radical-new-memory-tech-to-feed-rdna-3-5-graphics-and-zen-5-cpu-coresRead 15 October 2026AMD's page states a 256-bit LPDDR5x interface and no channel count. Counted in 64-bit channels, 256 bits is 4 channels. The same Tom's Hardware article quotes 256 GB/s of memory throughput, which fits 4 channels of 64 bits at 8000 MT/s.

21Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Max # of Memory Channels: 2.

Memory bandwidth (claimed)iPeak memory bandwidth the chip maker states. Not a measurement.

256GB/s4Press articletomshardware.com/pc-components/cpus/amds-beastly-strix-halo-ryzen-ai-max-debuts-with-radical-new-memory-tech-to-feed-rdna-3-5-graphics-and-zen-5-cpu-coresRead 15 October 2026Tom's Hardware, AMD's CES 2025 launch coverage: the Ryzen AI Max+ 395 delivers 256 GB/s of memory throughput. This is the maker's peak figure as reported, and it matches a 256-bit bus at LPDDR5X-8000 (256 bits / 8 x 8000 MT/s).

Not available

Max memoryiMost memory the chip maker says the processor supports. When the limit depends on the device, as for Apple chips, the highest limit the maker states for this chip in any device. The device goes in the source note. When Intel's documents differ, such as 192 GB in a launch brief and 256 GB on its spec page, the current spec page wins, since newer memory modules can raise the limit.

128GB1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Max. Memory 128 GB. AMD's launch release says up to 96 GB of it can be used by the graphics.

96GB1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Max Memory Size (dependent on memory type): 96 GB.

ECC memory supportiThe chip maker says the processor supports ECC memory. Most need a board that supports it too. DDR5 on-die ECC does not count.

Not available

Not available

Platform

5 of 7 differ

Socket or packageiDesktop socket, e.g. AM5 or LGA1851. Laptop chips are soldered, so this is the package the maker names, e.g. FP8. When the maker names several packages, in the socket row or the product IDs, list them all separated by commas, such as "FP7, FP7r2".

FP111Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026CPU Socket row reads FP11.

BGA25401Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Sockets Supported: FCBGA2540. Stored without the FC prefix.

Compatible chipsetsiDesktop chipsets the chip maker lists, with the brand and without the word chipset, e.g. AMD X870E, AMD B850 or Intel Z890. Same form as the motherboard chipset.

Not available

Not available

CPU PCIe lanesiUsable PCIe lanes from the processor, one row per PCIe generation. Lanes that link the CPU to the chipset are not counted. If the maker gives only a total, a review that lists the split may be used, and the source note names it. When the package also holds the chipset, as on Intel Raptor Lake U and H chips, ARK's total includes the chipset lanes. Those are not counted, so the datasheet split is used. On Intel Core Ultra chips, such as Panther Lake, the platform controller or SoC tile is part of the processor, so its lanes count.

AAMD Ryzen AI Max+ 395

PCIe generation	Lanes
4	16

Source1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026PCI Express Version is PCIe 4.0. Native PCIe Lanes (Total/Usable) is 16 and 16.

BIntel Core Ultra X9 388H

PCIe generation	Lanes
5	4
4	8

Source1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026PCI Express Revision 5.0 and 4.0. Max # of PCI Express Lanes 12. PCI Express Configurations: x4 PCIe5 (1x4, 2x2) and two rows of x4 PCIe4 (1x4, 2x2, 4x1). That gives 4 lanes at Gen 5 and 8 lanes at Gen 4, which adds up to the stated 12. The Quick Reference Guide gives the same split, 12 (8 G4, 4 G5). The Core Ultra 9 386H has 20 lanes with an extra 8 Gen 5 lanes.3Maker spec sheet (PDF)cdrdv2-public.intel.com/871380/Intel%20Core%20Ultra%20Series%203%20Processors%20-%20Quick%20Reference%20Guide%20v1.pdfRead 15 October 2026Page 3: Total Platform PCIe Lanes, CPU PCIe Config for the 388H reads 12 (8 G4, 4 G5).5Press articleold.chipsandcheese.com/2025/10/14/panther-lakes-reveal-at-itt-2025Read 15 October 2026The article says the largest configuration keeps the I/O configuration of the smallest one, which is 4 Gen 5 and 8 Gen 4 lanes for 12 lanes. The middle configuration, as in the Core Ultra 9 386H, adds 8 Gen 5 lanes.

CPU PCIe lanes in totaliUsable lanes from the processor. Lanes to the chipset are not counted.

16=CalculatedFrom CPU PCIe lanes

12=CalculatedFrom CPU PCIe lanes

Highest PCIe generationiHighest PCIe generation of the processor's own lanes. Chipset lanes are not counted.

4=CalculatedFrom CPU PCIe lanes

5=CalculatedFrom CPU PCIe lanes

Native USB portsiUSB ports the processor itself provides, as the chip maker lists them. The board or laptop decides which ones reach a connector. Chipset ports are not counted. eUSB2 is the embedded USB 2.0 link that Qualcomm lists. It needs a repeater on the board to become a USB 2.0 port.

AAMD Ryzen AI Max+ 395

Standard	Count
USB4	2
USB 3.2 Gen 2 (10 Gbps)	3
USB 2.0	3

Source1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Native USB 4 (40Gbps) 2, Native USB 3.2 Gen 2 (10Gbps) 3 and Native USB 2.0 (480Mbps) 3.

BIntel Core Ultra X9 388H

Standard	Count
Thunderbolt 4	4

Source1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Intel Thunderbolt 4: Yes. ARK gives no port count. The count of 4 comes from the Quick Reference Guide.3Maker spec sheet (PDF)cdrdv2-public.intel.com/871380/Intel%20Core%20Ultra%20Series%203%20Processors%20-%20Quick%20Reference%20Guide%20v1.pdfRead 15 October 2026Page 3: a merged cell for the 388H, 386H, 368H and 366H rows reads 4 Integrated Thunderbolt 4 Ports, Discrete Thunderbolt 5 Technology Support. Only the 4 integrated Thunderbolt 4 ports are stored. Thunderbolt 5 needs a separate controller on the laptop board.

Native USB4 or ThunderboltiThe processor itself provides USB4 or Thunderbolt. The board or laptop must still wire it to a port.

Yes=CalculatedFrom Native USB ports

Yes=CalculatedFrom Native USB ports

Integrated graphics

5 of 9 differ

Integrated graphicsiNo for chips that need a graphics card, such as Intel F models.

Yes1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026

Yes1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026

Integrated GPUiAs the chip maker names it.

AMD Radeon 8060S1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Graphics Model: Radeon 8060S Graphics, written AMD Radeon 8060S in AMD's launch release.

Intel Arc B3901Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026GPU Name: Intel Arc B390 GPU. Intel also sells an Arc Pro B390 name for vPro systems.

GPU coresiAs the chip maker counts them: AMD compute units, Intel Xe-cores, Apple GPU cores. Counts from different makers are not comparable. Intel execution units (EUs) are not Xe-cores and are not stored here. Apple sells some chips with two GPU core counts under one name, so it is per variant.

401Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Graphics Core Count 40.

121Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Xe-cores: 12. Ray Tracing: Yes.

Max GPU clockiIntel Graphics Max Dynamic Frequency or AMD Graphics Frequency.

2,900MHz1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Graphics Frequency 2900 MHz.

2,500MHz1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Graphics Max Dynamic Frequency 2.5 GHz.

Max displaysiDisplays the integrated GPU drives at the same time. The chip maker's total. Intel counts a built-in laptop panel in it. A count of external displays only, as Qualcomm gives, is left out and goes in the source note.

41Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Max Displays 4.

41Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026# of Displays Supported: 4. Max Resolution (DP) 7680 x 4320 at 60 Hz and eDP 3840 x 2400 at 120 Hz.

DisplayPort versioniHighest DisplayPort version the integrated GPU outputs, with the UHBR rate when stated.

2.1 UHBR201Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026DisplayPort Version 2.1 with the extensions Adaptive-Sync, HDR Metadata and UHBR20.

2.1 UHBR201Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Graphics Output: eDP 1.5, DP 2.1 UHBR20, HDMI 2.1 FRL.

HDMI versioniHighest HDMI version the integrated GPU outputs, as published.

2.11Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026HDMI Version 2.1.

2.1 FRL1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026Graphics Output: eDP 1.5, DP 2.1 UHBR20, HDMI 2.1 FRL.

Hardware video encodeiCodecs the chip maker states the integrated GPU encodes in hardware.

H.264, H.265 (HEVC), AV11Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Max Video Encode Bandwidth (SDR) lists H.264, H.265 and AV1.

H.264, H.265 (HEVC), AV11Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026H.264 and H.265 (HEVC) hardware encode and decode and AV1 encode and decode are listed. H.266 is decode only.

Hardware video decodeiCodecs the chip maker states the integrated GPU decodes in hardware.

H.264, H.265 (HEVC), VP9, AV11Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Max Video Decode Bandwidth lists H.264, VP9, H.265, AV1 and MJPEG. MJPEG is not an option in the schema.

H.264, H.265 (HEVC), AV1, H.266 (VVC)1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026H.264, H.265 (HEVC), AV1 and H.266 (decode only) are listed for hardware decode. ARK has no VP9 row, so VP9 is left out.

NPU

1 of 2 differ

NPU performanceiNPU only, as the chip maker states it. CPU and GPU AI throughput are not included. Precision, usually INT8, belongs in the source note.

50TOPS1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026NPU TOPS: Up to 50 TOPS. AMD's page names no precision.

50TOPS1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026NPU Peak TOPS (Int8): 50. ARK lists GPU Peak TOPS (Int8) 122 but has no Overall Peak TOPS row, so platform_tops is left out.

Total AI performanceiThe chip maker's combined AI figure for CPU, GPU and NPU together, e.g. Intel Overall Peak TOPS or AMD Overall TOPS.

126TOPS1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026Overall TOPS: Up to 126 TOPS. AMD's page names no precision. FP16 AI Compute is listed separately as up to 60 TFLOPS.

Not available

Measured performance

2 of 7 differ

Cinebench 2024 single-coreiMeasured by a reviewer, not claimed by the maker. Compare only results from the same reviewer.

Not available

Not available

Cinebench 2024 multi-coreiMeasured by a reviewer, not claimed by the maker. Compare only results from the same reviewer.

Not available

Not available

Cinebench 2026 single-threadiMeasured by a reviewer, not claimed by the maker. The CPU (Single Thread) test, not the CPU (Single Core) test with the core's second thread. Compare only results from the same reviewer.

Not available

Not available

Cinebench 2026 multi-coreiMeasured by a reviewer, not claimed by the maker. The CPU (Multiple Threads) test. Not comparable with Cinebench 2024. Compare only results from the same reviewer.

Not available

Not available

Geekbench 6 single-coreiMeasured by a reviewer, not claimed by the maker. Geekbench 6 only. Geekbench 7 has another scale and is not stored here. Intel Binary Optimization on or off goes in the source note. Store results with it off. Compare only results from the same reviewer.

Not available

3,036points7Reviewclub386.com/intel-core-ultra-series-3-panther-lake-reviewRead 15 October 2026Club386 Core Ultra Series 3 review of the X9 388H, 26 January 2026. App performance table: Geekbench 6 single-core 3,036. Tested in an Asus Zenbook 14 Duo (UX8407) with 32 GB LPDDR5X-9600, Windows Performance power mode and the Asus fan mode set to Performance, which the review says were used for all performance figures. The Geekbench 6 minor version is not stated and the review does not mention Intel Binary Optimization.

Geekbench 6 multi-coreiMeasured by a reviewer, not claimed by the maker. Geekbench 6 only. Geekbench 7 has another scale and is not stored here. Intel Binary Optimization on or off goes in the source note. Store results with it off. Compare only results from the same reviewer.

Not available

17,800points7Reviewclub386.com/intel-core-ultra-series-3-panther-lake-reviewRead 15 October 2026Geekbench 6 multi-core 17,800 in the same table. The review text calls this 62.9 percent faster than the Core Ultra 9 288V at 10,926, which matches. Same test laptop and settings. Tom's Hardware measured 3,031 and 17,283 in the same laptop and is not used, since Club386 is the Geekbench 6 reviewer for this type.

Game frame ratesiReviewer game results with this processor and the named graphics card, at native resolution with no upscaling and ray tracing off unless the source note says otherwise. Intel Binary Optimization or APO on or off goes in the source note. Store only results with it off. Compare only results from the same reviewer. A 99th percentile frame rate, as Tom's Hardware prints it, is stored as the 1 % low.

Not available

Not available

Measured power and temperature

1 of 5 differ

CPU power in multi-thread load (measured)iMeasured by a reviewer, not claimed by the maker. CPU package power in a sustained all-core load such as Cinebench multi-core. Not wall power. Compare only results from the same reviewer.

Not available

Not available

CPU power while gaming (measured)iMeasured by a reviewer, not claimed by the maker. Average CPU package power while gaming. Not wall power. Compare only results from the same reviewer. Intel Binary Optimization or APO on or off goes in the source note. A geomean that mixes games with it on and off is left out.

Not available

Not available

CPU power at idle (measured)iMeasured by a reviewer, not claimed by the maker. CPU package power at the reviewer's idle state. Not wall power. Compare only results from the same reviewer.

Not available

Not available

CPU temperature in stress testiMeasured by a reviewer, not claimed by the maker. Highest absolute CPU temperature in a sustained all-core load. A delta over room temperature is not stored. It depends on the cooler. Compare only results from the same reviewer.

Not available

Not available

All-core clock in stress testiMeasured by a reviewer, not claimed by the maker. Sustained clock under an all-core load. On chips with several core groups, the clock of the group with the highest rated clock, such as Intel P-cores, even when another group measures higher. Compare only results from the same reviewer.

Not available

2.86GHz8Measurementtomshardware.com/laptops/gaming-laptops/asus-zenbook-duo-2026-reviewRead 15 October 2026Tom's Hardware review of the Asus Zenbook Duo (UX8407) with the Core Ultra X9 388H, 26 January 2026. During its ten-run Cinebench 2026 stress test the performance cores ran at an average of 2.86 GHz, the efficiency cores at 2.7 GHz and the low-power efficient cores at 2.45 GHz. The P-cores are the fastest group, so 2.86 GHz is stored. The review does not state the laptop's power mode, fan profile or room temperature. The ten runs scored 3,872 on the first run and in the mid 3,800s after run 5, so no single Cinebench result is stored.

In the box

None of 2 known

Cooler in the boxiA cooler comes in the retail box. Tray versions have none.

Not available

Not available

Included cooler

Not available

Not available

Links

Not compared

Product page

www.amd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.html1Maker product pageamd.com/en/products/processors/laptop/ryzen/ai-300-series/amd-ryzen-ai-max-plus-395.htmlRead 15 October 2026

www.intel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.html1Maker product pageintel.com/content/www/us/en/products/sku/245526/intel-core-ultra-x9-processor-388h-18m-cache-up-to-5-10-ghz/specifications.htmlRead 15 October 2026

Spec sheet

Not available

cdrdv2-public.intel.com/871380/Intel%20Core%20Ultra%20Series%203%20Processors%20-%20Quick%20Reference%20Guide%20v1.pdf3Maker spec sheet (PDF)cdrdv2-public.intel.com/871380/Intel%20Core%20Ultra%20Series%203%20Processors%20-%20Quick%20Reference%20Guide%20v1.pdfRead 15 October 2026Intel Core Ultra Series 3 Processors Quick Reference Guide v1, SKU table on page 3 with a row for the Core Ultra X9 388H.

Manual

Not available

Not available

EPREL registration numberiId in the EU energy label database. Each size or model of a series has its own id, so it is per variant.

Not available

Not available

Warranty

None of 1 known

Manufacturer warranty (source market)iThe maker's warranty period for the exact SKU or configuration and market named in the source note. It applies only to that stated market, not every regional variant. Statutory consumer rights are excluded. The base term is stored. An extension that needs registration goes in the source note.

Not available

Not available

Both are also in the API. Back to all CPUs.