Verdict — Bigger With Just A Few More Features Than The NucBox K11
The NucBox K12 is larger than the NucBox K11 and so it naturally comes with a few more helpful features including an extra M.2 SSD slot, additional integrated ports, and a dedicated button for controlling the RGB LED. Its 120 W AC adapter is also smaller and much more compact than before.
Other than the changes mentioned above, however, the NucBox K12 performs roughly the same as its predecessor. It doesn't even offer integrated speakers or quieter fans despite the larger physical size. Serviceability could have also been made simpler.
Pros
Cons
Price and Availability
GMKtec is now shipping the NucBox K12 for $410 as a barebones unit or $550 when configured with 32 GB of RAM and 1 TB SSD. Alternatively, the model is on sale at Amazon for $520.
The NucBox K12 is an updated version of the NucBox K11 with both a newer AMD CPU and a refreshed chassis design. Much like its predecessor, the K12 targets professionals or business users who can exploit OcuLink and the wide range of integrated ports.
Our test unit comes with 32 GB of RAM and a 1 TB SSD There are no other CPU options at the moment beyond the Ryzen 7 H 255.
Competitors in this space include other professional mini PCs like the MSI Cubi series, Asus ExpertCenter series, or the Bosgame M series.
More GMKtec reviews:
Specifications
Case
The new design is noticeably larger and heavier than the outgoing NucBox K11 and many other mini PCs for that matter. In turn, The K12 utilizes a smaller 120 W AC adapter (~9.8 x 6.2 x 2.4 cm) plus additional internal storage options.
New to the model is the "Fan Mode" button on top which cycles through a preset number of RGB effects. Holding the button down turns off the RGB lights which addresses one of our biggest complaints about the RGB effects on the EVO-X1 and NucBox K11. Otherwise, the button has no impact on the performance of the system.
Connectivity
The model inherits all the ports found on the NucBox K11 including the OcuLink port plus additional USB-A and audio jack ports for good measure.
Communication
Networking | |
iperf3 transmit AXE11000 | |
GMK NucBox K12 | |
GMK EVO-X1 | |
Geekom A9 Max, AI 9 HX 370 | |
Minisforum EliteMini AI370 | |
GMK NucBox K11 | |
Beelink SER9 Pro | |
iperf3 receive AXE11000 | |
GMK EVO-X1 | |
Geekom A9 Max, AI 9 HX 370 | |
Minisforum EliteMini AI370 | |
Beelink SER9 Pro | |
GMK NucBox K12 | |
GMK NucBox K11 | |
iperf3 transmit AXE11000 6GHz | |
Geekom A9 Max, AI 9 HX 370 | |
GMK NucBox K12 | |
Minisforum EliteMini AI370 | |
iperf3 receive AXE11000 6GHz | |
Geekom A9 Max, AI 9 HX 370 | |
Minisforum EliteMini AI370 | |
GMK NucBox K12 |
Maintenance
Serviceability is poorly designed. While the four feet on the bottom are thumb screws, the grip is extremely poor and so the feet are difficult to remove. Even after removing these screws, four additional Phillips screws must be removed to access the motherboard. This is in contrast to most other mini PCs where servicing is much more straightforward.
The system supports up to three PCIe4 M.2 2280 SSDs (two PCIe4 x2 + one PCIe4 x4) up from two on the older NucBox K11 design. SODIMM remains user-upgradeable via 2x SODIMM slots.
Sustainability
The manufacturer makes no mention of recycled materials for manufacturing. The retail box consists of paper and cardboard with plastic wrapping.
Accessories And Warranty
There retail box includes a removable VESA mounting bracket and an HDMI cable. The usual one-year limited manufacturer warranty applies if purchased in the US.
Performance
Testing Conditions
We set Windows to the default Balanced mode prior to running the benchmarks below.
Processor
Despite the Ryzen 7 H 255 being relatively new, the CPU is still based on the last generation Zen 4 architecture with no integrated NPU and the older RDNA3 iGPU instead of RDNA3.5. Raw performance is therefore very close to the two-year old Ryzen 7 7840HS while being 20 to 25 percent slower than the Zen 5 Ryzen AI 9 HX 370. It is essentially a rebadged Ryzen 7 7840HS for all intents and purposes.
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
Cinebench R23 / Multi Core | |
GMK EVO-X2 | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
Asus NUC 14 Pro+ | |
Beelink SER8 | |
GMK NucBox K11 | |
Beelink SER9 Pro | |
Beelink GTI13 Ultra | |
Average AMD Ryzen 7 H 255 (15871 - 16402, n=2) | |
BOSGAME M1 | |
GMK NucBox K12 | |
Minisforum Venus Series NAB9 | |
HP Pro Mini 400 G9-937U0EA |
Cinebench R23 / Single Core | |
Beelink GTI13 Ultra | |
GMK EVO-X2 | |
GMK EVO-X1 | |
Minisforum EliteMini AI370 | |
Minisforum Venus Series NAB9 | |
HP Pro Mini 400 G9-937U0EA | |
GMK NucBox K11 | |
Beelink SER8 | |
Asus NUC 14 Pro+ | |
BOSGAME M1 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 (1697 - 1718, n=2) | |
GMK NucBox K12 |
Cinebench R20 / CPU (Multi Core) | |
GMK EVO-X2 | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
Asus NUC 14 Pro+ | |
Beelink SER9 Pro | |
GMK NucBox K11 | |
Beelink SER8 | |
Average AMD Ryzen 7 H 255 (6255 - 6480, n=2) | |
BOSGAME M1 | |
GMK NucBox K12 | |
Beelink GTI13 Ultra | |
Minisforum Venus Series NAB9 | |
HP Pro Mini 400 G9-937U0EA |
Cinebench R20 / CPU (Single Core) | |
Beelink GTI13 Ultra | |
GMK EVO-X2 | |
GMK EVO-X1 | |
Minisforum EliteMini AI370 | |
Minisforum Venus Series NAB9 | |
GMK NucBox K11 | |
HP Pro Mini 400 G9-937U0EA | |
Asus NUC 14 Pro+ | |
Beelink SER8 | |
BOSGAME M1 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 (660 - 677, n=2) | |
GMK NucBox K12 |
Cinebench R15 / CPU Multi 64Bit | |
GMK EVO-X2 | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
Asus NUC 14 Pro+ | |
Beelink SER8 | |
GMK NucBox K11 | |
BOSGAME M1 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 (2514 - 2588, n=2) | |
GMK NucBox K12 | |
Beelink GTI13 Ultra | |
Minisforum Venus Series NAB9 | |
HP Pro Mini 400 G9-937U0EA |
Cinebench R15 / CPU Single 64Bit | |
GMK EVO-X2 | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
Beelink GTI13 Ultra | |
GMK NucBox K11 | |
Beelink SER8 | |
Beelink SER9 Pro | |
BOSGAME M1 | |
Average AMD Ryzen 7 H 255 (266 - 276, n=2) | |
Minisforum Venus Series NAB9 | |
GMK NucBox K12 | |
Asus NUC 14 Pro+ | |
HP Pro Mini 400 G9-937U0EA |
Blender / v2.79 BMW27 CPU | |
HP Pro Mini 400 G9-937U0EA | |
Minisforum Venus Series NAB9 | |
GMK NucBox K12 | |
Average AMD Ryzen 7 H 255 (194 - 203, n=2) | |
Beelink GTI13 Ultra | |
BOSGAME M1 | |
Beelink SER9 Pro | |
GMK NucBox K11 | |
Beelink SER8 | |
Asus NUC 14 Pro+ | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
GMK EVO-X2 |
7-Zip 18.03 / 7z b 4 -mmt1 | |
Beelink SER8 | |
GMK NucBox K11 | |
BOSGAME M1 | |
Beelink GTI13 Ultra | |
GMK EVO-X2 | |
GMK NucBox K12 | |
Average AMD Ryzen 7 H 255 (6237 - 6261, n=2) | |
Beelink SER9 Pro | |
Minisforum EliteMini AI370 | |
Minisforum Venus Series NAB9 | |
GMK EVO-X1 | |
HP Pro Mini 400 G9-937U0EA | |
Asus NUC 14 Pro+ |
Geekbench 6.5 / Multi-Core | |
GMK EVO-X2 | |
Minisforum EliteMini AI370 | |
Asus NUC 14 Pro+ | |
Beelink SER8 | |
BOSGAME M1 | |
Beelink GTI13 Ultra | |
GMK NucBox K11 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 (12515 - 12825, n=2) | |
GMK EVO-X1 | |
GMK NucBox K12 | |
Minisforum Venus Series NAB9 | |
HP Pro Mini 400 G9-937U0EA |
Geekbench 6.5 / Single-Core | |
GMK EVO-X1 | |
Beelink GTI13 Ultra | |
GMK EVO-X2 | |
Minisforum EliteMini AI370 | |
GMK NucBox K11 | |
Beelink SER8 | |
BOSGAME M1 | |
Minisforum Venus Series NAB9 | |
HP Pro Mini 400 G9-937U0EA | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 (2545 - 2546, n=2) | |
GMK NucBox K12 | |
Asus NUC 14 Pro+ |
Geekbench 5.5 / Multi-Core | |
GMK EVO-X2 | |
Minisforum EliteMini AI370 | |
Asus NUC 14 Pro+ | |
Beelink GTI13 Ultra | |
GMK EVO-X1 | |
Beelink SER8 | |
BOSGAME M1 | |
GMK NucBox K11 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 (11604 - 11614, n=2) | |
GMK NucBox K12 | |
Minisforum Venus Series NAB9 | |
HP Pro Mini 400 G9-937U0EA |
Geekbench 5.5 / Single-Core | |
GMK EVO-X2 | |
GMK EVO-X1 | |
Beelink SER8 | |
GMK NucBox K11 | |
Beelink GTI13 Ultra | |
Minisforum EliteMini AI370 | |
BOSGAME M1 | |
GMK NucBox K12 | |
Minisforum Venus Series NAB9 | |
Average AMD Ryzen 7 H 255 (1810 - 1880, n=2) | |
HP Pro Mini 400 G9-937U0EA | |
Beelink SER9 Pro | |
Asus NUC 14 Pro+ |
HWBOT x265 Benchmark v2.2 / 4k Preset | |
GMK EVO-X2 | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
Beelink SER8 | |
GMK NucBox K11 | |
Beelink SER9 Pro | |
BOSGAME M1 | |
Average AMD Ryzen 7 H 255 (19.9 - 20.4, n=2) | |
GMK NucBox K12 | |
Asus NUC 14 Pro+ | |
Beelink GTI13 Ultra | |
Minisforum Venus Series NAB9 | |
HP Pro Mini 400 G9-937U0EA |
LibreOffice / 20 Documents To PDF | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
Asus NUC 14 Pro+ | |
GMK NucBox K12 | |
Average AMD Ryzen 7 H 255 (51 - 51.4, n=2) | |
Beelink SER9 Pro | |
GMK NucBox K11 | |
GMK EVO-X2 | |
HP Pro Mini 400 G9-937U0EA | |
BOSGAME M1 | |
Beelink SER8 | |
Beelink GTI13 Ultra | |
Minisforum Venus Series NAB9 |
R Benchmark 2.5 / Overall mean | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 (0.4587 - 0.503, n=2) | |
GMK NucBox K12 | |
Asus NUC 14 Pro+ | |
HP Pro Mini 400 G9-937U0EA | |
BOSGAME M1 | |
Minisforum Venus Series NAB9 | |
Beelink SER8 | |
GMK NucBox K11 | |
Minisforum EliteMini AI370 | |
Beelink GTI13 Ultra | |
GMK EVO-X1 | |
GMK EVO-X2 |
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
* ... 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 | |
Geekom A9 Max, AI 9 HX 370 | |
Minisforum EliteMini AI370 | |
GMK NucBox K11 | |
GMK EVO-X1 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 | |
GMK NucBox K12 |
AIDA64 / FP32 Ray-Trace | |
Geekom A9 Max, AI 9 HX 370 | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
GMK NucBox K11 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 (25404 - 26120, n=2) | |
GMK NucBox K12 |
AIDA64 / FPU Julia | |
Geekom A9 Max, AI 9 HX 370 | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
GMK NucBox K11 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 (104302 - 106590, n=2) | |
GMK NucBox K12 |
AIDA64 / CPU SHA3 | |
Geekom A9 Max, AI 9 HX 370 | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
GMK NucBox K11 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 (4604 - 4764, n=2) | |
GMK NucBox K12 |
AIDA64 / CPU Queen | |
Beelink SER9 Pro | |
GMK NucBox K11 | |
Average AMD Ryzen 7 H 255 (111673 - 115748, n=2) | |
GMK NucBox K12 | |
Geekom A9 Max, AI 9 HX 370 | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 |
AIDA64 / FPU SinJulia | |
Geekom A9 Max, AI 9 HX 370 | |
GMK EVO-X1 | |
Minisforum EliteMini AI370 | |
Beelink SER9 Pro | |
GMK NucBox K11 | |
Average AMD Ryzen 7 H 255 (14626 - 15123, n=2) | |
GMK NucBox K12 |
AIDA64 / FPU Mandel | |
Geekom A9 Max, AI 9 HX 370 | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
GMK NucBox K11 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 (56019 - 57378, n=2) | |
GMK NucBox K12 |
AIDA64 / CPU AES | |
GMK NucBox K11 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 (83999 - 86090, n=2) | |
GMK NucBox K12 | |
Minisforum EliteMini AI370 | |
Geekom A9 Max, AI 9 HX 370 | |
GMK EVO-X1 |
AIDA64 / CPU ZLib | |
Geekom A9 Max, AI 9 HX 370 | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
Beelink SER9 Pro | |
GMK NucBox K11 | |
Average AMD Ryzen 7 H 255 (1036 - 1066, n=2) | |
GMK NucBox K12 |
AIDA64 / FP64 Ray-Trace | |
Geekom A9 Max, AI 9 HX 370 | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
GMK NucBox K11 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 (13694 - 13941, n=2) | |
GMK NucBox K12 |
AIDA64 / CPU PhotoWorxx | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
Geekom A9 Max, AI 9 HX 370 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 (40115 - 41649, n=2) | |
GMK NucBox K11 | |
GMK NucBox K12 |
System Performance
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 | |
GMK NucBox K11 | |
GMK EVO-X1 | |
Geekom A9 Max, AI 9 HX 370 | |
Beelink SER9 Pro | |
Minisforum EliteMini AI370 | |
Average AMD Ryzen 7 H 255, AMD Radeon 780M | |
GMK NucBox K12 | |
GMK NucBox M7 Pro |
PCMark 10 / Score | |
Geekom A9 Max, AI 9 HX 370 | |
GMK NucBox K11 | |
GMK EVO-X1 | |
Minisforum EliteMini AI370 | |
GMK NucBox K12 | |
Average AMD Ryzen 7 H 255, AMD Radeon 780M (6907 - 6959, n=2) | |
Beelink SER9 Pro | |
GMK NucBox M7 Pro |
PCMark 10 / Essentials | |
GMK NucBox K11 | |
GMK EVO-X1 | |
Geekom A9 Max, AI 9 HX 370 | |
GMK NucBox K12 | |
GMK NucBox M7 Pro | |
Average AMD Ryzen 7 H 255, AMD Radeon 780M (10377 - 10502, n=2) | |
Beelink SER9 Pro | |
Minisforum EliteMini AI370 |
PCMark 10 / Productivity | |
GMK NucBox K11 | |
Geekom A9 Max, AI 9 HX 370 | |
GMK NucBox K12 | |
Minisforum EliteMini AI370 | |
Average AMD Ryzen 7 H 255, AMD Radeon 780M (9325 - 9440, n=2) | |
Beelink SER9 Pro | |
GMK EVO-X1 | |
GMK NucBox M7 Pro |
PCMark 10 / Digital Content Creation | |
Geekom A9 Max, AI 9 HX 370 | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
GMK NucBox K11 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255, AMD Radeon 780M (9227 - 9243, n=2) | |
GMK NucBox K12 | |
GMK NucBox M7 Pro |
CrossMark / Overall | |
Beelink SER9 Pro | |
Minisforum EliteMini AI370 | |
Average AMD Ryzen 7 H 255, AMD Radeon 780M (1769 - 1791, n=2) | |
GMK NucBox K11 | |
GMK NucBox K12 | |
Geekom A9 Max, AI 9 HX 370 | |
GMK EVO-X1 | |
GMK NucBox M7 Pro |
CrossMark / Productivity | |
Beelink SER9 Pro | |
GMK NucBox K11 | |
Average AMD Ryzen 7 H 255, AMD Radeon 780M (1709 - 1733, n=2) | |
GMK NucBox K12 | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
Geekom A9 Max, AI 9 HX 370 | |
GMK NucBox M7 Pro |
CrossMark / Creativity | |
Geekom A9 Max, AI 9 HX 370 | |
Minisforum EliteMini AI370 | |
GMK NucBox K11 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255, AMD Radeon 780M (1887 - 1895, n=2) | |
GMK NucBox K12 | |
GMK EVO-X1 | |
GMK NucBox M7 Pro |
CrossMark / Responsiveness | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255, AMD Radeon 780M (1617 - 1666, n=2) | |
GMK NucBox K12 | |
GMK NucBox K11 | |
GMK EVO-X1 | |
Minisforum EliteMini AI370 | |
Geekom A9 Max, AI 9 HX 370 | |
GMK NucBox M7 Pro |
WebXPRT 3 / Overall | |
GMK NucBox K11 | |
GMK EVO-X1 | |
GMK NucBox M7 Pro | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255, AMD Radeon 780M (290 - 300, n=2) | |
GMK NucBox K12 | |
Geekom A9 Max, AI 9 HX 370 | |
Minisforum EliteMini AI370 |
WebXPRT 4 / Overall | |
GMK NucBox K11 | |
GMK EVO-X1 | |
Beelink SER9 Pro | |
GMK NucBox K12 | |
Average AMD Ryzen 7 H 255, AMD Radeon 780M (245 - 245, n=2) | |
GMK NucBox M7 Pro | |
Minisforum EliteMini AI370 | |
Geekom A9 Max, AI 9 HX 370 |
Mozilla Kraken 1.1 / Total | |
GMK NucBox K12 | |
Average AMD Ryzen 7 H 255, AMD Radeon 780M (566 - 577, n=2) | |
GMK NucBox M7 Pro | |
Beelink SER9 Pro | |
Minisforum EliteMini AI370 | |
GMK NucBox K11 | |
Geekom A9 Max, AI 9 HX 370 | |
GMK EVO-X1 |
* ... smaller is better
PCMark 10 Score | 6959 points | |
Help |
AIDA64 / Memory Copy | |
GMK EVO-X1 | |
Minisforum EliteMini AI370 | |
Geekom A9 Max, AI 9 HX 370 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 (67673 - 74184, n=2) | |
GMK NucBox K11 | |
GMK NucBox K12 |
AIDA64 / Memory Read | |
GMK EVO-X1 | |
Minisforum EliteMini AI370 | |
Geekom A9 Max, AI 9 HX 370 | |
GMK NucBox K11 | |
GMK NucBox K12 | |
Average AMD Ryzen 7 H 255 (58305 - 61129, n=2) | |
Beelink SER9 Pro |
AIDA64 / Memory Write | |
Beelink SER9 Pro | |
GMK EVO-X1 | |
Minisforum EliteMini AI370 | |
Average AMD Ryzen 7 H 255 (85497 - 106026, n=2) | |
GMK NucBox K11 | |
Geekom A9 Max, AI 9 HX 370 | |
GMK NucBox K12 |
AIDA64 / Memory Latency | |
GMK EVO-X1 | |
Minisforum EliteMini AI370 | |
Beelink SER9 Pro | |
Average AMD Ryzen 7 H 255 (93.7 - 103.9, n=2) | |
GMK NucBox K12 | |
GMK NucBox K11 | |
Geekom A9 Max, AI 9 HX 370 |
* ... smaller is better
DPC Latency
DPC Latencies / LatencyMon - interrupt to process latency (max), Web, Youtube, Prime95 | |
Minisforum EliteMini AI370 | |
Geekom A9 Max, AI 9 HX 370 | |
GMK NucBox K12 | |
GMK EVO-X1 | |
Beelink SER9 Pro | |
GMK NucBox K11 |
* ... smaller is better
Storage Devices
Our unit ships with the same 1 TB Crucial P3 Plus CT1000P3PSSD8 NVMe SSD as found on many other mini PCs including the Beelink SER9 Pro or Minisforum EliteMini AI370.
Drive Performance Rating - Percent | |
Geekom A9 Max, AI 9 HX 370 | |
Minisforum EliteMini AI370 | |
GMK NucBox K11 | |
GMK EVO-X1 | |
Beelink SER9 Pro | |
GMK NucBox K12 |
* ... smaller is better
Disk Throttling: DiskSpd Read Loop, Queue Depth 8
GPU Performance
The integrated Radeon 780M is approximately 10 to 15 percent faster than the older Radeon 680M while being 10 to 15 percent slower than the newer Radeon 890M. It is nonetheless powerful enough to run the latest titles at 900p to 1080p albeit at lower settings.
3DMark 11 Performance | 13189 points | |
3DMark Cloud Gate Standard Score | 37614 points | |
3DMark Fire Strike Score | 7755 points | |
3DMark Time Spy Score | 3215 points | |
Help |
* ... smaller is better
Performance Rating - Percent | |
Minisforum EliteMini AI370 -1! | |
Geekom A9 Max, AI 9 HX 370 -1! | |
GMK EVO-X1 | |
GMK NucBox K11 | |
Beelink SER9 Pro | |
GMK NucBox K12 |
Cyberpunk 2077 2.2 Phantom Liberty - 1920x1080 Ultra Preset (FSR off) | |
Minisforum EliteMini AI370 | |
Geekom A9 Max, AI 9 HX 370 | |
GMK EVO-X1 | |
GMK NucBox K11 | |
GMK NucBox K12 | |
Beelink SER9 Pro |
Baldur's Gate 3 - 1920x1080 Ultra Preset AA:T | |
Geekom A9 Max, AI 9 HX 370 | |
Minisforum EliteMini AI370 | |
GMK NucBox K11 | |
GMK EVO-X1 | |
Beelink SER9 Pro | |
GMK NucBox K12 |
GTA V - 1920x1080 Highest Settings possible AA:4xMSAA + FX AF:16x | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
Geekom A9 Max, AI 9 HX 370 | |
Beelink SER9 Pro | |
GMK NucBox K11 | |
GMK NucBox K12 |
Final Fantasy XV Benchmark - 1920x1080 High Quality | |
Minisforum EliteMini AI370 | |
GMK EVO-X1 | |
Geekom A9 Max, AI 9 HX 370 | |
Beelink SER9 Pro | |
GMK NucBox K11 | |
GMK NucBox K12 |
Strange Brigade - 1920x1080 ultra AA:ultra AF:16 | |
GMK EVO-X1 | |
Beelink SER9 Pro | |
GMK NucBox K11 | |
GMK NucBox K12 |
Dota 2 Reborn - 1920x1080 ultra (3/3) best looking | |
GMK NucBox K11 | |
Minisforum EliteMini AI370 | |
Geekom A9 Max, AI 9 HX 370 | |
GMK EVO-X1 | |
GMK NucBox K12 | |
Beelink SER9 Pro |
X-Plane 11.11 - 1920x1080 high (fps_test=3) | |
Minisforum EliteMini AI370 | |
Geekom A9 Max, AI 9 HX 370 | |
GMK EVO-X1 | |
GMK NucBox K11 | |
GMK NucBox K12 | |
Beelink SER9 Pro |
F1 24 | |
1920x1080 Low Preset AA:T AF:16x | |
GMK EVO-X2 | |
Minisforum EliteMini AI370 | |
Beelink SER9 Pro | |
GMK NucBox K12 | |
Average AMD Radeon 780M (43.3 - 94.8, n=25) | |
Asus NUC 14 Pro+ | |
HP Pro Mini 400 G9-937U0EA | |
1920x1080 Medium Preset AA:T AF:16x | |
GMK EVO-X2 | |
Minisforum EliteMini AI370 | |
Beelink SER9 Pro | |
GMK NucBox K12 | |
Average AMD Radeon 780M (36.4 - 80.2, n=26) | |
Asus NUC 14 Pro+ | |
HP Pro Mini 400 G9-937U0EA | |
1920x1080 High Preset AA:T AF:16x | |
GMK EVO-X2 | |
Minisforum EliteMini AI370 | |
Beelink SER9 Pro | |
GMK NucBox K12 | |
Average AMD Radeon 780M (23.3 - 59.9, n=26) | |
Asus NUC 14 Pro+ | |
HP Pro Mini 400 G9-937U0EA | |
1920x1080 Ultra High Preset AA:T AF:16x | |
GMK EVO-X2 | |
Beelink SER9 Pro | |
GMK NucBox K12 | |
Average AMD Radeon 780M (6.39 - 14.5, n=16) |
Cyberpunk 2077 ultra FPS Chart
low | med. | high | ultra | |
---|---|---|---|---|
GTA V (2015) | 164.9 | 154.4 | 65 | 27.8 |
Dota 2 Reborn (2015) | 154 | 108.9 | 93.2 | 81.5 |
Final Fantasy XV Benchmark (2018) | 78.4 | 39.1 | 27.6 | |
X-Plane 11.11 (2018) | 61.1 | 45.3 | 38 | |
Strange Brigade (2018) | 168.5 | 67.1 | 56.5 | 49.5 |
Baldur's Gate 3 (2023) | 39.9 | 32.9 | 27 | 26.7 |
Cyberpunk 2077 2.2 Phantom Liberty (2023) | 43.5 | 33.9 | 27.1 | 23.6 |
F1 24 (2024) | 83.4 | 73.3 | 55.4 | 12.3 |
Emissions
System Noise
The cooling solution consists of two fans and a vapor chamber cooler. Fan noise is nonetheless on the loud side at 35 dB(A) against a silent background of 24 dB(A) even when idling on desktop. Noise levels top out at the low 40 dB(A) range when higher loads are introduced to be average for a mini PC.
Noise Level
Idle |
| 35.4 / 35.4 / 39 dB(A) |
Load |
| 42.4 / 41.4 dB(A) |
![]() | ||
30 dB silent 40 dB(A) audible 50 dB(A) loud |
||
min: ![]() ![]() ![]() |
GMK NucBox K12 Radeon 780M, Ryzen 7 H 255, Crucial P3 Plus CT1000P3PSSD8 | GMK NucBox K11 Radeon 780M, R9 8945HS, Crucial P3 Plus CT1000P3PSSD8 | GMK EVO-X1 Radeon 890M, Ryzen AI 9 HX 370, Crucial P3 Plus CT1000P3PSSD8 | Beelink SER9 Pro Radeon 780M, Ryzen 7 H 255, Crucial P3 Plus CT1000P3PSSD8 | Geekom A9 Max, AI 9 HX 370 Radeon 890M, Ryzen AI 9 HX 370, Lexar NQ7A1 | Minisforum EliteMini AI370 Radeon 890M, Ryzen AI 9 HX 370, Crucial P3 Plus CT1000P3PSSD8 | |
---|---|---|---|---|---|---|
Noise | 13% | 5% | 15% | 4% | 6% | |
off / environment * (dB) | 23.5 | 24 -2% | 23.3 1% | 23 2% | 24.8 -6% | 24.8 -6% |
Idle Minimum * (dB) | 35.4 | 25.2 29% | 27.8 21% | 29.1 18% | 25.4 28% | 25 29% |
Idle Average * (dB) | 35.4 | 30.2 15% | 30.3 14% | 30.1 15% | 25.4 28% | 25.1 29% |
Idle Maximum * (dB) | 39 | 30.2 23% | 30.4 22% | 31.7 19% | 25.5 35% | 29.1 25% |
Load Average * (dB) | 42.4 | 36.4 14% | 44.6 -5% | 33.7 21% | 49.8 -17% | 47 -11% |
Cyberpunk 2077 ultra * (dB) | 40.4 | 36.6 9% | 44.2 -9% | 33.8 16% | 49.6 -23% | 44 -9% |
Load Maximum * (dB) | 41.4 | 39 6% | 46.1 -11% | 35.2 15% | 49.8 -20% | 47.1 -14% |
Witcher 3 ultra * (dB) | 49.7 |
* ... smaller is better
Temperature
Stress Test
The CPU would stabilize at 3.9 GHz, 54 W, and 78 C when initiating Prime95 stress. When running Cyberpunk 2077 to represent typical gaming loads, the integrated GPU would stabilize at 2550 MHz and 70 C which is close to what we recorded on the smaller NucBox K11.
Average CPU Clock (GHz) | GPU Clock (MHz) | Average CPU Temperature (°C) | |
System Idle | -- | 800 | 36 |
Prime95 Stress | 3.9 | -- | 78 |
Prime95 + FurMark Stress | 3.9 | 2171 | 70 |
Cyberpunk 2077 Stress | 1.1 | 2550 | 70 |
Energy Management
Power Consumption
Performance-per-watt is decent for a Zen 4 CPU but otherwise relatively poor when compared to newer Zen 5 options. The GMK EVO-X1, for example, is simultaneously faster and less demanding to be a significantly more power efficient option.
Off / Standby | ![]() ![]() |
Idle | ![]() ![]() ![]() |
Load |
![]() ![]() |
![]() ![]() ![]() ![]() ![]() ![]() ![]() | |
Key:
min: ![]() ![]() ![]() |
GMK NucBox K12 Ryzen 7 H 255, Radeon 780M, Crucial P3 Plus CT1000P3PSSD8, , x, | GMK NucBox K11 R9 8945HS, Radeon 780M, Crucial P3 Plus CT1000P3PSSD8, , x, | GMK EVO-X1 Ryzen AI 9 HX 370, Radeon 890M, Crucial P3 Plus CT1000P3PSSD8, , x, | Beelink SER9 Pro Ryzen 7 H 255, Radeon 780M, Crucial P3 Plus CT1000P3PSSD8, , x, | Geekom A9 Max, AI 9 HX 370 Ryzen AI 9 HX 370, Radeon 890M, Lexar NQ7A1, , x, | Minisforum EliteMini AI370 Ryzen AI 9 HX 370, Radeon 890M, Crucial P3 Plus CT1000P3PSSD8, , x, | |
---|---|---|---|---|---|---|
Power Consumption | 9% | 29% | 22% | 26% | 32% | |
Idle Minimum * (Watt) | 9.9 | 6.6 33% | 5.4 45% | 7 29% | 2.4 76% | 3.4 66% |
Idle Average * (Watt) | 10.1 | 9.1 10% | 5.8 43% | 7.1 30% | 4.6 54% | 4.9 51% |
Idle Maximum * (Watt) | 13.1 | 11.3 14% | 5.8 56% | 7.5 43% | 9.2 30% | 7.4 44% |
Load Average * (Watt) | 69.8 | 76.8 -10% | 72.1 -3% | 62.3 11% | 78.4 -12% | 70.7 -1% |
Cyberpunk 2077 ultra external monitor * (Watt) | 86.3 | 83 4% | 72.5 16% | 80.1 7% | 77.8 10% | 69.6 19% |
Load Maximum * (Watt) | 92.8 | 88 5% | 76.1 18% | 82.9 11% | 95.1 -2% | 82.6 11% |
* ... smaller is better
Power Consumption Cyberpunk / Stress Test
Power Consumption external Monitor
Notebookcheck Rating
The NucBox K12 introduces more features over the NucBox K11, but raw performance remains almost the same.
GMK NucBox K12
- 09/26/2025 v8
Allen Ngo
Potential Competitors in Comparison
Image | Model / Review | Price | Weight | Height | Display |
---|---|---|---|---|---|
GMK NucBox K12 AMD Ryzen 7 H 255 ⎘ AMD Radeon 780M ⎘ 32 GB Memory, 1024 GB SSD | Amazon: $749.99 | 938 g | 78 mm | x | |
GMK NucBox K11 AMD Ryzen 9 8945HS ⎘ AMD Radeon 780M ⎘ 32 GB Memory, 1024 GB SSD | Amazon: $849.99 List Price: 600 USD | 630 g | 58 mm | x | |
GMK EVO-X1 AMD Ryzen AI 9 HX 370 ⎘ AMD Radeon 890M ⎘ 32 GB Memory, 1024 GB SSD | Amazon: $1,049.99 List Price: 900 USD | 595 g | 63.2 mm | x | |
Beelink SER9 Pro AMD Ryzen 7 H 255 ⎘ AMD Radeon 780M ⎘ 32 GB Memory, 1024 GB SSD | Amazon: $649.00 List Price: 520 USD | 812 g | 44.7 mm | x | |
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 | |
Minisforum EliteMini AI370 AMD Ryzen AI 9 HX 370 ⎘ AMD Radeon 890M ⎘ 32 GB Memory, 1024 GB SSD | Amazon: 1. $44.99 PowerHOOD 19V AC/DC Adapter ... 2. $47.28 Kircuit 19V AC/DC Adapter fo... 3. $48.78 Onerbl AC-DC Adapter for Cha... List Price: 1499 EUR | 472 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. We never accept compensation or payment in return for our reviews. 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.