Verdict — Mini-ITX PC powered by mobile 120 W AMD Zen 5 CPU
The Framework Desktop is unique in that it's a mini-ITX form factor powered by a 120 W mobile AMD Ryzen Zen 5 CPU. Smaller mini PCs with this same processor are available for similar prices including the GMK EVO-X2 or BOSGAME M5, but the Framework runs quieter, cooler, and surprisingly more efficiently by adequate margins while being easier to service or repair.
There is still room for improvement. Installing the Expansion Cards, for example, requires lifting or tilting the unit off the table. Waste heat exits from the side panel which can affect positioning. There is no support for PCIe5 SSDs, upgradeable memory, AMD PRO CPUs or ECC RAM, but these are likely limitations of the AMD processor and not the Framework design. The setbacks are otherwise minor for a solid mini PC/desktop hybrid solution.
Pros
Cons
Price and Availability
The base configuration with the Ryzen AI Max 385 starts at $1100 while our review unit as configured starts at $1600.
Unveiled in February of this year, the Framework Desktop is the company's third major product and first Mini-ITX desktop in the Framework family. The system carries the same design philosophies as the Laptop 13.5 or Laptop 16 series including ease of repairability and compatibility with Expansion Cards as top priorities.
Our test unit is a middle configuration with the AMD Ryzen AI Max+ 395 CPU, integrated Radeon 8060S GPU, and 64 GB of RAM for approximately $1600. Lower-end options ship with the Ryzen AI Max 385 and Radeon 8050S instead while the highest-end option includes 128 GB of soldered RAM.
As a mini PC and desktop hybrid, the Framework Desktop competes with models like the GMK EVO-X2, BOSGAME M5, or Beelink SER9 Pro all of which offer the same processor option as the Framework Desktop.
More Framework reviews:
Specifications
Case
The Framework Desktop is larger and heavier than other PCs with the exact same processor including the GMK EVO-X2, BOSGAME M5, or Beelink GTR9 Pro. Of course, this discrepancy is due to the superior serviceability and integrated PSU of the Framework design which negates the need for external AC adapters.
Connectivity
Notable ports include the five video-out ports along the rear plus two USB-C 3.2 Gen. 2 ports on the front. The two front ports are compatible with the same Expansion Cards that have been available since the original Framework Laptop in 2021.
Communication
| Networking | |
| iperf3 transmit AXE11000 | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| HP Z2 Mini G1a | |
| Framework Desktop Ryzen AI Max | |
| Geekom A9 Max, AI 9 HX 370 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| iperf3 receive AXE11000 | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| HP Z2 Mini G1a | |
| Framework Desktop Ryzen AI Max | |
| iperf3 transmit AXE11000 6GHz | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Geekom A9 Max, AI 9 HX 370 | |
| HP Z2 Mini G1a | |
| Framework Desktop Ryzen AI Max | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| iperf3 receive AXE11000 6GHz | |
| BOSGAME M5 | |
| HP Z2 Mini G1a | |
| GMK EVO-X2 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| Framework Desktop Ryzen AI Max | |
Maintenance
The motherboard is accessible without any tools as the top and side panels can slide off without much hassle. Users have direct access to two M.2 2280 SSD slots, a removable WLAN module, and a free PCIe x4 expansion slot.
Sustainability
The system is manufactured with 30 percent PCC plastic and 50 percent PCC aluminum while the packaging is fully recyclable and plastic-free. Nonetheless, accessories are individually packaged including each Expansion Card and Tile leading to an overuse of cardboard if ordering a handful of accessories at once.
Accessories And Warranty
The DIY Edition includes the CPU fan mounting kit and screwdriver for installation while the Expansion Cards, Tiles, and CPU fan are all optional. The plastic Tiles adorn the front panel of the system in a 3 x 7 grid purely for aesthetic and personalization purposes. Framework could perhaps support RGB LED Tiles in the future for an extra level of customization.
The standard one-year limited manufacturer warranty applies if purchased in the US.
Performance
Testing Conditions
We set Windows to Performance mode prior to running the benchmarks below. VRAM is set to 8 GB by default on the Framework Desktop, but we've set this to 16 GB for the purposes of this review.
Processor
Raw multi-thread performance with the Ryzen AI Max+ 395 is comparable to the Core Ultra 9 275HX otherwise known as one of Intel's fastest Arrow Lake-HX options at the moment.
Cinebench R15 Multi Loop
Cinebench R23: Multi Core | Single Core
Cinebench R20: CPU (Multi Core) | CPU (Single Core)
Cinebench R15: CPU Multi 64Bit | CPU Single 64Bit
Blender: v2.79 BMW27 CPU
7-Zip 18.03: 7z b 4 | 7z b 4 -mmt1
Geekbench 6.5: Multi-Core | Single-Core
Geekbench 5.5: Multi-Core | Single-Core
HWBOT x265 Benchmark v2.2: 4k Preset
LibreOffice : 20 Documents To PDF
R Benchmark 2.5: Overall mean
| CPU Performance rating | |
| Alienware 16X Aurora, RTX 5070 | |
| Average AMD Ryzen AI Max+ 395 | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Framework Desktop Ryzen AI Max | |
| HP Z2 Mini G1a | |
| Zotac GeForce RTX 5060 Low Profile -11! | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| Beelink SER9 Pro | |
| HP OmniStudio X 32-c0077ng -1! | |
| Cinebench R23 / Multi Core | |
| HP Z2 Mini G1a | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Alienware 16X Aurora, RTX 5070 | |
| Framework Desktop Ryzen AI Max | |
| Average AMD Ryzen AI Max+ 395 (29063 - 37423, n=7) | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| Beelink SER9 Pro | |
| HP OmniStudio X 32-c0077ng | |
| Cinebench R23 / Single Core | |
| Alienware 16X Aurora, RTX 5070 | |
| Average AMD Ryzen AI Max+ 395 (2034 - 2052, n=7) | |
| Framework Desktop Ryzen AI Max | |
| GMK EVO-X2 | |
| BOSGAME M5 | |
| HP Z2 Mini G1a | |
| Geekom A9 Max, AI 9 HX 370 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| HP OmniStudio X 32-c0077ng | |
| Beelink SER9 Pro | |
| Cinebench R20 / CPU (Multi Core) | |
| HP Z2 Mini G1a | |
| BOSGAME M5 | |
| Alienware 16X Aurora, RTX 5070 | |
| GMK EVO-X2 | |
| Framework Desktop Ryzen AI Max | |
| Average AMD Ryzen AI Max+ 395 (11357 - 14641, n=7) | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| Beelink SER9 Pro | |
| HP OmniStudio X 32-c0077ng | |
| Cinebench R20 / CPU (Single Core) | |
| Alienware 16X Aurora, RTX 5070 | |
| BOSGAME M5 | |
| Framework Desktop Ryzen AI Max | |
| Average AMD Ryzen AI Max+ 395 (791 - 803, n=7) | |
| GMK EVO-X2 | |
| HP Z2 Mini G1a | |
| Geekom A9 Max, AI 9 HX 370 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Beelink SER9 Pro | |
| HP OmniStudio X 32-c0077ng | |
| Cinebench R15 / CPU Multi 64Bit | |
| BOSGAME M5 | |
| HP Z2 Mini G1a | |
| GMK EVO-X2 | |
| Alienware 16X Aurora, RTX 5070 | |
| Average AMD Ryzen AI Max+ 395 (4739 - 5881, n=7) | |
| Framework Desktop Ryzen AI Max | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| HP OmniStudio X 32-c0077ng | |
| Beelink SER9 Pro | |
| Cinebench R15 / CPU Single 64Bit | |
| Alienware 16X Aurora, RTX 5070 | |
| BOSGAME M5 | |
| Framework Desktop Ryzen AI Max | |
| Average AMD Ryzen AI Max+ 395 (314 - 318, n=7) | |
| GMK EVO-X2 | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| Beelink SER9 Pro | |
| HP OmniStudio X 32-c0077ng | |
| Blender / v2.79 BMW27 CPU | |
| Beelink SER9 Pro | |
| Geekom A9 Max, AI 9 HX 370 | |
| Zotac GeForce RTX 5060 Low Profile | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Alienware 16X Aurora, RTX 5070 | |
| Framework Desktop Ryzen AI Max | |
| Average AMD Ryzen AI Max+ 395 (86 - 107, n=5) | |
| HP Z2 Mini G1a | |
| GMK EVO-X2 | |
| BOSGAME M5 | |
| 7-Zip 18.03 / 7z b 4 | |
| BOSGAME M5 | |
| HP Z2 Mini G1a | |
| Average AMD Ryzen AI Max+ 395 (116237 - 136872, n=5) | |
| GMK EVO-X2 | |
| Framework Desktop Ryzen AI Max | |
| Alienware 16X Aurora, RTX 5070 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| Beelink SER9 Pro | |
| HP OmniStudio X 32-c0077ng | |
| 7-Zip 18.03 / 7z b 4 -mmt1 | |
| Alienware 16X Aurora, RTX 5070 | |
| Average AMD Ryzen AI Max+ 395 (6474 - 6540, n=5) | |
| Framework Desktop Ryzen AI Max | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| HP Z2 Mini G1a | |
| Geekom A9 Max, AI 9 HX 370 | |
| Beelink SER9 Pro | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| HP OmniStudio X 32-c0077ng | |
| Geekbench 6.5 / Single-Core | |
| Zotac GeForce RTX 5060 Low Profile | |
| Alienware 16X Aurora, RTX 5070 | |
| Framework Desktop Ryzen AI Max | |
| Average AMD Ryzen AI Max+ 395 (2791 - 2978, n=5) | |
| HP Z2 Mini G1a | |
| GMK EVO-X2 | |
| Geekom A9 Max, AI 9 HX 370 | |
| BOSGAME M5 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Beelink SER9 Pro | |
| HP OmniStudio X 32-c0077ng | |
| Geekbench 5.5 / Single-Core | |
| Zotac GeForce RTX 5060 Low Profile | |
| Alienware 16X Aurora, RTX 5070 | |
| GMK EVO-X2 | |
| Average AMD Ryzen AI Max+ 395 (2184 - 2275, n=5) | |
| HP Z2 Mini G1a | |
| Framework Desktop Ryzen AI Max | |
| BOSGAME M5 | |
| Geekom A9 Max, AI 9 HX 370 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Beelink SER9 Pro | |
| HP OmniStudio X 32-c0077ng | |
| HWBOT x265 Benchmark v2.2 / 4k Preset | |
| Alienware 16X Aurora, RTX 5070 | |
| Average AMD Ryzen AI Max+ 395 (32.4 - 40.3, n=5) | |
| BOSGAME M5 | |
| Framework Desktop Ryzen AI Max | |
| GMK EVO-X2 | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| Beelink SER9 Pro | |
| HP OmniStudio X 32-c0077ng | |
| LibreOffice / 20 Documents To PDF | |
| HP OmniStudio X 32-c0077ng | |
| HP Z2 Mini G1a | |
| Geekom A9 Max, AI 9 HX 370 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| BOSGAME M5 | |
| Beelink SER9 Pro | |
| Average AMD Ryzen AI Max+ 395 (45.8 - 58.4, n=5) | |
| GMK EVO-X2 | |
| Framework Desktop Ryzen AI Max | |
| Alienware 16X Aurora, RTX 5070 | |
| R Benchmark 2.5 / Overall mean | |
| Beelink SER9 Pro | |
| HP OmniStudio X 32-c0077ng | |
| Alienware 16X Aurora, RTX 5070 | |
| BOSGAME M5 | |
| Geekom A9 Max, AI 9 HX 370 | |
| Average AMD Ryzen AI Max+ 395 (0.3984 - 0.4094, n=5) | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| HP Z2 Mini G1a | |
| GMK EVO-X2 | |
| Framework Desktop Ryzen AI Max | |
* ... smaller is better
AIDA64: FP32 Ray-Trace | FPU Julia | CPU SHA3 | CPU Queen | FPU SinJulia | FPU Mandel | CPU AES | CPU ZLib | FP64 Ray-Trace | CPU PhotoWorxx
| Performance rating | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Framework Desktop Ryzen AI Max -1! | |
| HP Z2 Mini G1a | |
| Average AMD Ryzen AI Max+ 395 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| AIDA64 / FP32 Ray-Trace | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Average AMD Ryzen AI Max+ 395 (64158 - 82725, n=5) | |
| Framework Desktop Ryzen AI Max | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| AIDA64 / FPU Julia | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Average AMD Ryzen AI Max+ 395 (183760 - 227429, n=5) | |
| Framework Desktop Ryzen AI Max | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| AIDA64 / CPU SHA3 | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Framework Desktop Ryzen AI Max | |
| Average AMD Ryzen AI Max+ 395 (8151 - 9475, n=5) | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| AIDA64 / CPU Queen | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Framework Desktop Ryzen AI Max | |
| HP Z2 Mini G1a | |
| Average AMD Ryzen AI Max+ 395 (145339 - 154404, n=5) | |
| Geekom A9 Max, AI 9 HX 370 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| AIDA64 / FPU SinJulia | |
| Framework Desktop Ryzen AI Max | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Average AMD Ryzen AI Max+ 395 (29155 - 31931, n=5) | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| AIDA64 / FPU Mandel | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Average AMD Ryzen AI Max+ 395 (97193 - 121827, n=5) | |
| Framework Desktop Ryzen AI Max | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| AIDA64 / CPU AES | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Average AMD Ryzen AI Max+ 395 (139734 - 158617, n=4) | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| AIDA64 / CPU ZLib | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| HP Z2 Mini G1a | |
| Framework Desktop Ryzen AI Max | |
| Average AMD Ryzen AI Max+ 395 (0 - 2431, n=5) | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| AIDA64 / FP64 Ray-Trace | |
| GMK EVO-X2 | |
| BOSGAME M5 | |
| HP Z2 Mini G1a | |
| Average AMD Ryzen AI Max+ 395 (34297 - 43587, n=5) | |
| Framework Desktop Ryzen AI Max | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| AIDA64 / CPU PhotoWorxx | |
| Framework Desktop Ryzen AI Max | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| HP Z2 Mini G1a | |
| Average AMD Ryzen AI Max+ 395 (81626 - 94259, n=5) | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
PCMark 10: Score | Essentials | Productivity | Digital Content Creation
CrossMark: Overall | Productivity | Creativity | Responsiveness
WebXPRT 3: Overall
WebXPRT 4: Overall
Mozilla Kraken 1.1: Total
| Performance rating | |
| BOSGAME M5 | |
| Average AMD Ryzen AI Max+ 395, AMD Radeon 8060S | |
| Framework Desktop Ryzen AI Max | |
| GMK EVO-X2 | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| PCMark 10 / Score | |
| GMK EVO-X2 | |
| Average AMD Ryzen AI Max+ 395, AMD Radeon 8060S (8732 - 9125, n=5) | |
| Framework Desktop Ryzen AI Max | |
| BOSGAME M5 | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| PCMark 10 / Essentials | |
| GMK EVO-X2 | |
| Framework Desktop Ryzen AI Max | |
| Average AMD Ryzen AI Max+ 395, AMD Radeon 8060S (10949 - 12005, n=5) | |
| BOSGAME M5 | |
| HP Z2 Mini G1a | |
| Geekom A9 Max, AI 9 HX 370 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| PCMark 10 / Productivity | |
| GMK EVO-X2 | |
| Framework Desktop Ryzen AI Max | |
| Average AMD Ryzen AI Max+ 395, AMD Radeon 8060S (9762 - 10377, n=5) | |
| BOSGAME M5 | |
| HP Z2 Mini G1a | |
| Geekom A9 Max, AI 9 HX 370 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| PCMark 10 / Digital Content Creation | |
| Average AMD Ryzen AI Max+ 395, AMD Radeon 8060S (16185 - 16755, n=5) | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Framework Desktop Ryzen AI Max | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| CrossMark / Overall | |
| Average AMD Ryzen AI Max+ 395, AMD Radeon 8060S (1948 - 2274, n=5) | |
| BOSGAME M5 | |
| Framework Desktop Ryzen AI Max | |
| GMK EVO-X2 | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| CrossMark / Productivity | |
| BOSGAME M5 | |
| Average AMD Ryzen AI Max+ 395, AMD Radeon 8060S (1701 - 2790, n=5) | |
| Framework Desktop Ryzen AI Max | |
| GMK EVO-X2 | |
| HP Z2 Mini G1a | |
| Geekom A9 Max, AI 9 HX 370 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| CrossMark / Creativity | |
| Average AMD Ryzen AI Max+ 395, AMD Radeon 8060S (2331 - 2949, n=5) | |
| BOSGAME M5 | |
| HP Z2 Mini G1a | |
| GMK EVO-X2 | |
| Framework Desktop Ryzen AI Max | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| CrossMark / Responsiveness | |
| Average AMD Ryzen AI Max+ 395, AMD Radeon 8060S (1541 - 1719, n=5) | |
| BOSGAME M5 | |
| Framework Desktop Ryzen AI Max | |
| GMK EVO-X2 | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| WebXPRT 3 / Overall | |
| HP Z2 Mini G1a | |
| Framework Desktop Ryzen AI Max | |
| Average AMD Ryzen AI Max+ 395, AMD Radeon 8060S (302 - 358, n=5) | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Geekom A9 Max, AI 9 HX 370 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| WebXPRT 4 / Overall | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| HP Z2 Mini G1a | |
| Average AMD Ryzen AI Max+ 395, AMD Radeon 8060S (252 - 317, n=5) | |
| Framework Desktop Ryzen AI Max | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Geekom A9 Max, AI 9 HX 370 | |
| Mozilla Kraken 1.1 / Total | |
| GMK EVO-X2 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Average AMD Ryzen AI Max+ 395, AMD Radeon 8060S (424 - 453, n=5) | |
| Geekom A9 Max, AI 9 HX 370 | |
| HP Z2 Mini G1a | |
| BOSGAME M5 | |
| Framework Desktop Ryzen AI Max | |
* ... smaller is better
| PCMark 10 Score | 8875 points | |
Help | ||
| AIDA64 / Memory Copy | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Average AMD Ryzen AI Max+ 395 (148904 - 163031, n=5) | |
| Framework Desktop Ryzen AI Max | |
| HP Z2 Mini G1a | |
| GMK EVO-X2 | |
| BOSGAME M5 | |
| Geekom A9 Max, AI 9 HX 370 | |
| AIDA64 / Memory Read | |
| GMK EVO-X2 | |
| BOSGAME M5 | |
| Framework Desktop Ryzen AI Max | |
| Average AMD Ryzen AI Max+ 395 (121931 - 123769, n=5) | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| AIDA64 / Memory Write | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Average AMD Ryzen AI Max+ 395 (209816 - 216896, n=5) | |
| BOSGAME M5 | |
| Framework Desktop Ryzen AI Max | |
| GMK EVO-X2 | |
| Geekom A9 Max, AI 9 HX 370 | |
| AIDA64 / Memory Latency | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| HP Z2 Mini G1a | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Average AMD Ryzen AI Max+ 395 (131.9 - 140.5, n=5) | |
| Framework Desktop Ryzen AI Max | |
| Geekom A9 Max, AI 9 HX 370 | |
* ... smaller is better
| DPC Latencies / LatencyMon - interrupt to process latency (max), Web, Youtube, Prime95 | |
| Framework Desktop Ryzen AI Max | |
| HP Z2 Mini G1a | |
| Geekom A9 Max, AI 9 HX 370 | |
| GMK EVO-X2 | |
| BOSGAME M5 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
* ... smaller is better
| Drive Performance rating - Percent | |
| GMK EVO-X2 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| HP Z2 Mini G1a | |
| Framework Desktop Ryzen AI Max | |
| Geekom A9 Max, AI 9 HX 370 | |
| BOSGAME M5 | |
* ... smaller is better
Disk Throttling: DiskSpd Read Loop, Queue Depth 8
GPU Performance
Raw graphics performance with the Radeon 8060S lies in between the mobile GeForce RTX 4060 and mobile RTX 4070 according to 3DMark DX11 and DX12 tests. The Nvidia GPUs are nonetheless more effective at real-time ray tracing effects as evident by their much higher Speed Way and Port Royal scores.
| 3DMark 11 Performance | 33216 points | |
| 3DMark Cloud Gate Standard Score | 67013 points | |
| 3DMark Fire Strike Score | 26308 points | |
| 3DMark Time Spy Score | 11002 points | |
| 3DMark Steel Nomad Score | 2164 points | |
Help | ||
* ... smaller is better
Gaming Performance
Although not primarily a gaming machine, overall performance is nonetheless comparable to the mobile GeForce RTX 4060 to be a decent 1080p gaming machine. Unfortunately, the system is not compatible with the recent AMD FSR4 update as the GPU is based on the same RDNA 3.5 architecture as the Radeon 890M and not the Radeon RX 9000 RDNA4 series.
| Performance rating - Percent | |
| GMK EVO-X2 | |
| BOSGAME M5 -1! | |
| HP Z2 Mini G1a | |
| Framework Desktop Ryzen AI Max | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 -1! | |
| Cyberpunk 2077 2.2 Phantom Liberty - 1920x1080 Ultra Preset (FSR off) | |
| HP Z2 Mini G1a | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Framework Desktop Ryzen AI Max | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| Baldur's Gate 3 - 1920x1080 Ultra Preset AA:T | |
| BOSGAME M5 | |
| Framework Desktop Ryzen AI Max | |
| GMK EVO-X2 | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| GTA V - 1920x1080 Highest Settings possible AA:4xMSAA + FX AF:16x | |
| GMK EVO-X2 | |
| BOSGAME M5 | |
| Framework Desktop Ryzen AI Max | |
| HP Z2 Mini G1a | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| Final Fantasy XV Benchmark - 1920x1080 High Quality | |
| GMK EVO-X2 | |
| BOSGAME M5 | |
| HP Z2 Mini G1a | |
| Framework Desktop Ryzen AI Max | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| Strange Brigade - 1920x1080 ultra AA:ultra AF:16 | |
| GMK EVO-X2 | |
| BOSGAME M5 | |
| HP Z2 Mini G1a | |
| Framework Desktop Ryzen AI Max | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Dota 2 Reborn - 1920x1080 ultra (3/3) best looking | |
| Framework Desktop Ryzen AI Max | |
| GMK EVO-X2 | |
| BOSGAME M5 | |
| HP Z2 Mini G1a | |
| Geekom A9 Max, AI 9 HX 370 | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| X-Plane 11.11 - 1920x1080 high (fps_test=3) | |
| HP Z2 Mini G1a | |
| GMK EVO-X2 | |
| Framework Desktop Ryzen AI Max | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| Geekom A9 Max, AI 9 HX 370 | |
| Indiana Jones and the Great Circle | |
| 1920x1080 High Preset | |
| Zotac GeForce RTX 5060 Low Profile | |
| GMK EVO-X2 | |
| HP Z2 Mini G1a | |
| Average AMD Radeon 8060S (49.4 - 65.8, n=4) | |
| Framework Desktop Ryzen AI Max | |
| Alienware 16X Aurora, RTX 5070 | |
| 1920x1080 Very Ultra Preset | |
| Zotac GeForce RTX 5060 Low Profile | |
| HP Z2 Mini G1a | |
| GMK EVO-X2 | |
| Average AMD Radeon 8060S (40.3 - 64.6, n=4) | |
| Framework Desktop Ryzen AI Max | |
| Alienware 16X Aurora, RTX 5070 | |
| 2560x1440 Very Ultra + FSR Quality | |
| GMK EVO-X2 | |
| Average AMD Radeon 8060S (48.4 - 54.7, n=3) | |
| HP Z2 Mini G1a | |
| Framework Desktop Ryzen AI Max | |
| 2560x1440 Very Ultra Preset | |
| HP Z2 Mini G1a | |
| Zotac GeForce RTX 5060 Low Profile | |
| GMK EVO-X2 | |
| Average AMD Radeon 8060S (27.4 - 57.6, n=4) | |
| Framework Desktop Ryzen AI Max | |
| Alienware 16X Aurora, RTX 5070 | |
| Assassin's Creed Shadows | |
| 1920x1080 High | |
| Alienware 16X Aurora, RTX 5070 | |
| Zotac GeForce RTX 5060 Low Profile | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Framework Desktop Ryzen AI Max | |
| Average AMD Radeon 8060S (47 - 57, n=4) | |
| HP ZBook Ultra G1a A3ZQ0ET | |
| 1920x1080 Ultra High | |
| Zotac GeForce RTX 5060 Low Profile | |
| BOSGAME M5 | |
| GMK EVO-X2 | |
| Framework Desktop Ryzen AI Max | |
| Alienware 16X Aurora, RTX 5070 | |
| Average AMD Radeon 8060S (32 - 40, n=4) | |
| 2560x1440 Ultra High | |
| Alienware 16X Aurora, RTX 5070 | |
| Zotac GeForce RTX 5060 Low Profile | |
| GMK EVO-X2 | |
| BOSGAME M5 | |
| Framework Desktop Ryzen AI Max | |
| Average AMD Radeon 8060S (25 - 34, n=4) | |
| 2560x1440 Ultra High + DLSS Quality | |
| Zotac GeForce RTX 5060 Low Profile | |
| Alienware 16X Aurora, RTX 5070 | |
| 2560x1440 Ultra High + FSR Quality | |
| Framework Desktop Ryzen AI Max | |
| BOSGAME M5 | |
| Average AMD Radeon 8060S (23 - 40, n=4) | |
| GMK EVO-X2 | |
Cyberpunk 2077 ultra FPS Chart
| low | med. | high | ultra | QHD FSR | QHD | 4K | |
|---|---|---|---|---|---|---|---|
| GTA V (2015) | 185.7 | 182.3 | 181.5 | 97 | 75.7 | ||
| Dota 2 Reborn (2015) | 228 | 205 | 171.1 | 147.5 | 99.4 | ||
| Final Fantasy XV Benchmark (2018) | 204 | 132.3 | 93.3 | 67.8 | |||
| X-Plane 11.11 (2018) | 93 | 72.9 | 58.3 | ||||
| Strange Brigade (2018) | 566 | 260 | 215 | 187 | 122.3 | ||
| Baldur's Gate 3 (2023) | 146 | 121.5 | 102.9 | 100.8 | 71.8 | ||
| Cyberpunk 2077 2.2 Phantom Liberty (2023) | 152.1 | 124.4 | 97.7 | 86.1 | 51.4 | ||
| Alan Wake 2 (2023) | 76.2 | 67 | 64.8 | 42.6 | |||
| F1 24 (2024) | 269 | 248 | 187.5 | 58.1 | 61.9 | 38 | |
| Black Myth: Wukong (2024) | 87 | 67 | 51 | 28 | 29 | 21 | |
| Indiana Jones and the Great Circle (2024) | 76.7 | 58.9 | 54.1 | 48.4 | 33.4 | ||
| Monster Hunter Wilds (2025) | 71.5 | 61.7 | 49.1 | 35.6 | |||
| Assassin's Creed Shadows (2025) | 72 | 67 | 57 | 40 | 40 | 33 |
Emissions
System Noise
Fan noise is almost inaudible under most low to low-medium loads. Perhaps most importantly, RPM pulsing is kept to a minimum in contrast to smaller mini PCs where fan speeds tend to change more frequently and unexpectedly.
Noise level
| Idle |
| 23.4 / 23.4 / 23.4 dB(A) |
| Load |
| 36.6 / 40.1 dB(A) |
![]() | ||
30 dB silent 40 dB(A) audible 50 dB(A) loud |
||
min: | ||
| Framework Desktop Ryzen AI Max Radeon 8060S, Ryzen AI Max+ 395, WD PC SN7100S SDFPMSL-1T00-1101 | GMK EVO-X2 Radeon 8060S, Ryzen AI Max+ 395, Crucial P3 Plus CT1000P3PSSD8 | BOSGAME M5 Radeon 8060S, Ryzen AI Max+ 395, Kingston OM8PGP42048N-A0 | HP Z2 Mini G1a Radeon 8060S, Ryzen AI Max+ PRO 395, Samsung PM9A1 MZVL22T0HBLB | HP ZBook Ultra G1a A3ZQ0ET Radeon 8050S, Ryzen AI Max PRO 390, SK hynix PC801 HFS001TEJ9X101N | Geekom A9 Max, AI 9 HX 370 Radeon 890M, Ryzen AI 9 HX 370, Lexar NQ7A1 | |
|---|---|---|---|---|---|---|
| Noise | -18% | -14% | -38% | -8% | -17% | |
| off / environment * (dB) | 23.2 | 24.9 -7% | 24.8 -7% | 24.9 -7% | 24.6 -6% | 24.8 -7% |
| Idle Minimum * (dB) | 23.4 | 27.4 -17% | 25 -7% | 32.1 -37% | 24.6 -5% | 25.4 -9% |
| Idle Average * (dB) | 23.4 | 27.5 -18% | 25.1 -7% | 32.4 -38% | 24.6 -5% | 25.4 -9% |
| Idle Maximum * (dB) | 23.4 | 27.5 -18% | 25.1 -7% | 32.6 -39% | 24.6 -5% | 25.5 -9% |
| Load Average * (dB) | 36.6 | 47.5 -30% | 48.8 -33% | 41.1 -12% | 40.7 -11% | 49.8 -36% |
| Cyberpunk 2077 ultra * (dB) | 39.8 | 47.3 -19% | 46.6 -17% | 67 -68% | 45.4 -14% | 49.6 -25% |
| Load Maximum * (dB) | 40.1 | 47.2 -18% | 48.6 -21% | 67.3 -68% | 45.4 -13% | 49.8 -24% |
| Witcher 3 ultra * (dB) | 47.4 | 46.5 | 70 | 49.7 |
* ... smaller is better
Temperature
Stress Test
| Average CPU Clock (GHz) | GPU Clock (MHz) | Average CPU Temperature (°C) | |
| System Idle | -- | -- | 28 |
| Prime95 Stress | 3.8 | -- | 64 |
| Prime95 + FurMark Stress | 2.7 | 1129 | 62 |
| Cyberpunk 2077 Stress | 0.7 | 2873 | 52 |
Energy Management
Power Consumption
The Framework Desktop is more efficient than the GMK EVO-X2 even though both come equipped with the same Ryzen AI Max Ryzen AI Max+ 395 processor for the same level of performance. Running Prime95 or Cyberpunk 2077 would demand ~144 W each compared to ~178 W on the GMK system as shown by the comparison chart below.
| Off / Standby | |
| Idle | |
| Load |
|
Key:
min: | |
| Framework Desktop Ryzen AI Max Ryzen AI Max+ 395, Radeon 8060S, WD PC SN7100S SDFPMSL-1T00-1101, , x, | GMK EVO-X2 Ryzen AI Max+ 395, Radeon 8060S, Crucial P3 Plus CT1000P3PSSD8, , x, | BOSGAME M5 Ryzen AI Max+ 395, Radeon 8060S, Kingston OM8PGP42048N-A0, , x, | HP Z2 Mini G1a Ryzen AI Max+ PRO 395, Radeon 8060S, Samsung PM9A1 MZVL22T0HBLB, , x, | HP ZBook Ultra G1a A3ZQ0ET Ryzen AI Max PRO 390, Radeon 8050S, SK hynix PC801 HFS001TEJ9X101N, IPS LED, 1920x1200, 14" | Geekom A9 Max, AI 9 HX 370 Ryzen AI 9 HX 370, Radeon 890M, Lexar NQ7A1, , x, | |
|---|---|---|---|---|---|---|
| Power Consumption | -0% | -17% | -12% | 26% | 40% | |
| Idle Minimum * (Watt) | 10 | 6.5 35% | 9.3 7% | 7.1 29% | 4.5 55% | 2.4 76% |
| Idle Average * (Watt) | 10.5 | 8.2 22% | 11.5 -10% | 9.5 10% | 7.5 29% | 4.6 56% |
| Idle Maximum * (Watt) | 10.9 | 10.5 4% | 16.4 -50% | 13.8 -27% | 8.3 24% | 9.2 16% |
| Load Average * (Watt) | 81.1 | 84.3 -4% | 83.9 -3% | 76.7 5% | 80.7 -0% | 78.4 3% |
| Cyberpunk 2077 ultra external monitor * (Watt) | 140.6 | 178.1 -27% | 174.2 -24% | 195.1 -39% | 104.1 26% | 77.8 45% |
| Load Maximum * (Watt) | 164.2 | 214.9 -31% | 200.9 -22% | 245.4 -49% | 125.1 24% | 95.1 42% |
| Cyberpunk 2077 ultra * (Watt) | 106.5 |
* ... smaller is better
Power Consumption Cyberpunk / Stress Test
Power Consumption external Monitor
Notebookcheck Rating
A Mini-ITX desktop powered by an unconventional mobile AMD Zen 5 CPU for a versatile system.
Framework Desktop Ryzen AI Max
- 09/17/2025 v8
Allen Ngo
Potential Competitors in Comparison
Image | Model / Review | Price | Weight | Height | Display |
|---|---|---|---|---|---|
| Framework Desktop Ryzen AI Max AMD Ryzen AI Max+ 395 ⎘ AMD Radeon 8060S ⎘ 64 GB Memory, 1024 GB SSD | Amazon: 1. $1,549.95 GPD Win Max 2 2025 [AMD Ryze... List Price: 1600 USD | 3.1 kg | 96.8 mm | x | |
| GMK EVO-X2 AMD Ryzen AI Max+ 395 ⎘ AMD Radeon 8060S ⎘ 64 GB Memory, 1024 GB SSD | Amazon: $1,699.99 List Price: 1499€ | 1.6 kg | 77 mm | x | |
| BOSGAME M5 AMD Ryzen AI Max+ 395 ⎘ AMD Radeon 8060S ⎘ 128 GB Memory, 2024 GB SSD | Amazon: 1. $21.99 UL Listed 19V 3A 3.42A AC Ad... 2. $23.65 Mini PC Computer Protective ... List Price: 1469€ | 1.3 kg | 57 mm | x | |
| HP Z2 Mini G1a AMD Ryzen AI Max+ PRO 395 ⎘ AMD Radeon 8060S ⎘ 64 GB Memory, 2048 GB SSD | Amazon: 1. $1,964.76 HP Z2 Mini G1a Workstation -... 2. $3,551.36 HP Z2 Mini G1a Workstation -... 3. $2,315.52 HP Z2 Mini G1a Workstation -... List Price: 3175 EUR | 2.6 kg | 85.5 mm | x | |
| HP ZBook Ultra G1a A3ZQ0ET AMD Ryzen AI Max PRO 390 ⎘ AMD Radeon 8050S ⎘ 32 GB Memory, 1024 GB SSD | Amazon: $2,510.10 List Price: 2545 Euro | 1.6 kg | 19 mm | 14.00" 1920x1200 162 PPI IPS LED | |
| Geekom A9 Max, AI 9 HX 370 AMD Ryzen AI 9 HX 370 ⎘ AMD Radeon 890M ⎘ 32 GB Memory, 2048 GB SSD | Amazon: $999.00 List Price: 999€ | 709 g | 46.9 mm | x |
Transparency
The selection of devices to be reviewed is made by our editorial team. The test sample was provided to the author as a loan by the manufacturer or retailer for the purpose of this review. The lender had no influence on this review, nor did the manufacturer receive a copy of this review before publication. There was no obligation to publish this review. As an independent media company, Notebookcheck is not subjected to the authority of manufacturers, retailers or publishers.
This is how Notebookcheck is testing
Every year, Notebookcheck independently reviews hundreds of laptops and smartphones using standardized procedures to ensure that all results are comparable. We have continuously developed our test methods for around 20 years and set industry standards in the process. In our test labs, high-quality measuring equipment is utilized by experienced technicians and editors. These tests involve a multi-stage validation process. Our complex rating system is based on hundreds of well-founded measurements and benchmarks, which maintains objectivity. Further information on our test methods can be found here.


















































