The mini PC market in mid-2026 has separated into two news cycles that almost never share a paragraph. On one side, the press is busy with Intel’s new 18A budget chip and the Wildcat Lake mini PCs we covered last week: tiny boxes in the $179 to $349 band, designed to displace the N100 and N305 as the default sub-$300 pick. On the other side, the AMD Strix Halo coverage keeps escalating, with the Ryzen AI Max+ 395 mini PCs we wrote up in our local LLM inference guide selling for $1,800 to $2,200 and earning their headlines on the strength of a 128 GB unified-memory party trick most readers will never use.
What almost nobody is writing about is the $500 to $750 band that sits between them. The boxes in that band run the same Ryzen 7 8845HS that powers half the “best mini PC” listicles two years running, pair it with the Radeon 780M iGPU, fit 32 GB of DDR5 and a 1 TB NVMe drive in the case, idle below 10 watts of system power, and stay below 40 dB even under a Cinebench loop. They are the single most boring news beat in the category, which is why the press skips them, and they are the right answer for most readers reading this.
The measured ladder, not the spec sheet
The case for the mid-tier is easier to make with three numbers than with three paragraphs. The Intel N100 that anchors the budget tier scores around 5,327 on the PassMark CPU Mark chart. The Ryzen 7 8845HS that anchors the mid-tier scores around 28,378 on the same chart. The Ryzen AI Max+ 395 that anchors the premium tier scores in the 49,000 to 55,000 range, per PassMark’s listing (the sample size has been climbing fast since launch and the chart’s current reading sits near the upper end of that band). Read those out loud once, then look at the prices, and the value math is already done: the mid-tier is roughly 2.3 cents per CPU Mark at $650 direct, the budget tier is roughly 4.7 cents per point, and the premium tier sits around 3.3 to 3.6 cents per point depending on which PassMark snapshot you take.
Cinebench R23 multi-core tells the same story with fewer sceptics. NotebookCheck’s measured run on the Beelink SER8 put the 8845HS at 16,661 points in the test config; Hostbor’s UM880 Plus run hit 17,070 on the same silicon, with single-core landing at 1,787. An Intel N100 lands around 2,500 in the same test, an N305 lands around 5,000, and the premium-tier numbers for the AI Max+ 395 land roughly twice as high as the mid-tier multi-core score. Roughly six-times the N100, half the premium tier, in one Cinebench loop.
The iGPU story is the surprise. The Radeon 780M built into every 8845HS chip typically lands between 2,700 and 3,600 in 3DMark Time Spy across mini PC implementations, depending heavily on memory speed and chassis cooling. That puts the integrated GPU at GTX 1650 Ti territory, several times what the N100’s UHD iGPU manages, and inside earshot of a six-year-old midrange discrete card. For 1080p gaming, light Blender previews, ScreenSpace shaders, and most non-CUDA workloads, the iGPU stops being the bottleneck.
The power-draw story is the one that earns this tier its homelab fans. ServeTheHome’s SER8 review measured idle system power at 7 to 10 watts, with package (CPU-internal) power sitting between 3.5 and 4.7 watts. Push it under a sustained load and the SER8 climbs to 77 to 79 watts of system power, with the chip holding 54 to 58 watts of package power across the entire run rather than the 60-second burst-and-throttle pattern that defines most mini PCs in the same shape. Noise stayed inside a 36 to 39 dB envelope, with the studio’s own noise floor at 34 dB. The reviewer summary was that the SER8 is one of the few mini PCs in this category that does not throttle, which is exactly the property you want from a box that is going to spend three years in a closet running Proxmox.
The spec stack worth caring about
The four products this article anchors on share more than they differ. Every one of them runs an 8-core, 16-thread Zen 4 Ryzen chip, the Radeon 780M with 12 RDNA 3 compute units, 32 GB of DDR5-5600 SODIMMs (not soldered LPDDR5, which is the property that makes the rest of this paragraph possible), a 1 TB PCIe 4.0 NVMe drive in an M.2 slot, 2.5 Gigabit Ethernet, at least one USB4 port at 40 Gbps, and modern Wi-Fi (Wi-Fi 6 on the Beelink SER8, Wi-Fi 6E on the rest). The four boxes diverge on chassis, OCuLink versus pure USB4 for external GPU, single versus dual M.2 slots (the SER8 and UM880 Plus carry two, the Geekom A8 and A8 Max carry one), single versus dual Ethernet, and whether the lid pops with a fingernail or a screwdriver.
Beelink SER8
- +Sustained 54W to 58W package power with no throttling
- +7W to 10W system idle
- +dual-channel SODIMM
- +single USB4 port
- +mature firmware
- -No OCuLink
- -only one 2.5GbE NIC
- -plastic chassis
- -single-fan cooling under the sustained load
The DDR5 SODIMM choice matters more than the spec sheet lets on. Soldered LPDDR5X has slightly higher memory bandwidth and a real benefit for the 780M iGPU’s gaming performance, but the SODIMM stacks across this tier give you the upgrade path: 32 GB now, 96 GB in three years when DDR5 has fallen another 30 percent, without throwing the chassis away. For a single-developer workstation that is going to host four Electron apps, a JVM, Docker, and a Postgres container at the same time, 96 GB of upgrade headroom is a feature.
MINISFORUM UM880 Plus
- +OCuLink at this price (via the included M.2-adapter board)
- +8K display output via USB4
- +sub-35 dB noise floor
- +USB4 plus 2.5GbE on the rear
- -OCuLink uses one of the two M.2 slots (you cannot run dual NVMe and an eGPU simultaneously)
- -eGPU on Linux is finicky
- -larger chassis than the SER8
- -street price moves a lot (Hostbor's review tracked $349 to $650-plus depending on the week)
OCuLink is the tier’s quiet feature, and it deserves a reality check before anyone buys the UM880 Plus on the strength of it. Minisforum ships OCuLink as an M.2-adapter board rather than a native rear-panel port, which means turning it on costs one of the two M.2 storage slots: dual NVMe and an eGPU are not a simultaneous option. An OCuLink eGPU enclosure plus a midrange dGPU (a refurb RTX 4060 or a new 4070, say) lands the all-in cost around $1,100 to $1,400, which has crept into the lower premium tier. On Linux, OCuLink eGPU support depends on the kernel version, the BIOS, and the chassis manufacturer’s choices around PCIe hotplug. Phoronix’s OCuLink and eGPU coverage has tracked the patchset for two years now; it works, but the it-works-on-Friday-but-not-Monday energy is still in the air for Ubuntu 22.04. If a plug-and-play discrete GPU on day one matters more than the long-term upgrade path, the answer is a slightly larger chassis with a real desktop GPU slot, not OCuLink.
GEEKOM A8
- +Smaller chassis than the SER8 or UM880 Plus
- +premium aluminium top plate
- +dual SODIMM and a single Gen 4 NVMe slot in the smallest footprint of the tier
- -Single 2.5GbE NIC
- -no OCuLink
- -fan curve more aggressive than the SER8 under sustained load
For the Ryzen 9 8945HS variants like the Geekom A8, the practical buyer’s read is straightforward. The 8945HS clocks 100 MHz higher on boost than the 8845HS (5.2 GHz vs 5.1 GHz) and 200 MHz higher on base (4.0 GHz vs 3.8 GHz), but it shares the same Hawk Point silicon, the same Radeon 780M, the same TDP envelope, and within margin of error the same measured Cinebench numbers across published reviews. Treat the 8945HS as a rebadge with a slightly higher SKU price; the buying advice does not change. The Geekom A8 Max variant tracked here ships with the Ryzen 7 8745HS instead (the same 8845HS silicon with the NPU disabled): pick it for chassis and port preference, not for a CPU step-up.
Three readers who want this tier
The single-developer workstation case is the easiest. A modern web stack with VS Code, Node, Docker, Postgres in a container, Slack, two browsers with 40 tabs apiece, and a Zoom call in the background lives comfortably inside the 8845HS power envelope and the 32 GB of DDR5. The single-thread responsiveness is good enough that nothing in the toolchain feels laggy, and the multi-thread headroom handles a npm test run or a TypeScript compile without the laptop fans spinning up. The cost-per-CPU-Mark advantage compounds over a four-year working life: a $650 mini PC depreciates differently than a $1,800 laptop.
The family-desktop replacement is the second. The Radeon 780M plays Stardew Valley, Hades, and the entire indie back catalogue without breaking a sweat; Helldivers 2 runs at 1080p on low-to-medium presets in the 35 to 50 fps range that reviewers consistently report, not at max settings; older AAA titles whose minimum spec lists “GTX 1650” run at 1080p medium with FSR doing some of the work. This is not a flagship GPU and it will not pretend to be one. It runs Plex transcodes for the household. It does video calls without choking. It runs a Steam library and a Disney+ stream without picking a side. The household never has to think about the box again.
The homelab single-box host is the third, and this is where the power-draw measurement earns its keep. A $650 SER8 running 24x7 at the ServeTheHome-measured 8 watt idle draws about 70 kWh of electricity a year. At twelve cents a kilowatt-hour, that is roughly eight dollars and forty cents in annual power, the cost of two coffees. The same workload on an old tower idling at 60 watts costs about $63 a year. Over four years the mini PC wins back $220 in electricity alone, which is a meaningful fraction of its purchase price.
The Linux story is the one homelab readers will want to read carefully. Recent kernels (6.5+) include solid amdgpu support for the Radeon 780M; most distros boot cleanly out of the box, though a minority of users still need a nomodeset first boot before the driver settles in. Proxmox 8.2 and later books straight through; Docker, K3s, and the standard self-hosted-services stack all run as expected. The catch is the 32 GB memory ceiling for the under-$700 SKUs: it is enough for the homelab and self-hosted-services workload the self-hosting renaissance piece describes, and it is enough for local LLM inference on quantized 7B and 13B parameter models, but it is not enough for the 70B-class models that the Strix Halo tier exists to run.
Where mid-tier is the wrong answer
The mid-tier is not the answer for everyone, and pretending it is would be the same listicle hedge this article opens by rejecting. Three readers should look elsewhere.
The first is anyone running serious local LLM inference on models above the 13B parameter class. The 780M is a perfectly capable iGPU, but BIOS allocation typically caps the iGPU at roughly half the system RAM as VRAM (around 16 GB on a 32 GB box, around 32 GB on a 64 GB box, with exact numbers varying by firmware). For 70B-parameter inference at usable token rates, the Ryzen AI Max+ 395’s 128 GB of unified memory and the 256 GB/s bandwidth it ships with is genuinely the right tool for the job, and our Strix Halo local LLM guide goes deep on why. If the workload is LLM inference, do not save money in the wrong place.
The second is anyone whose work depends on CUDA, real discrete-GPU gaming, or video editing pipelines that lean on NVENC. The 780M will not run DaVinci Resolve at the speed an RTX 4070 will, and no amount of clever benchmarking will close the gap. If the day-job tool needs a real dGPU, a slightly larger chassis with a desktop GPU slot (a NZXT H2 Mini ITX-class build or comparable) is the right shape for the money. OCuLink can bridge the gap; a desktop GPU does not need bridging.
The third is the reader who already owns a discrete-GPU desktop and is just shopping for a small always-on box to host the smart-home stack or run a private VPN. The mid-tier overshoots that workload by a wide margin, and a Beelink EQ12 or similar budget-tier N100 box at $179 to $249 handles it without the cost-per-CPU-Mark math changing the answer. The mid-tier is the answer when the box is doing real work; it is the wrong answer when the box is a permanent background process.
The buying read
The four mini PCs this article centers on are the four to keep on the shortlist: Beelink SER8 at $650 direct (closer to $900 at Newegg this week), Minisforum UM880 Plus anywhere from $349 to roughly $650-plus depending on the promotion, Geekom A8 at roughly the same band as the SER8 with the 8945HS rebadge on its top SKU, and Geekom A8 Max for the 8745HS variant if a slightly different chassis and port mix is the deciding factor. The street prices in this article are accurate as of late May 2026; with DRAM and NVMe pricing the way it has been since the DRAM-spike piece ran in April, expect the band to drift by $50 to $100 in either direction across the rest of the year.
The press will keep covering the news cycles either side of this tier, because launch news is what generates headlines. That is the inefficiency this article exists to point at: the most boring corner of the mini PC market is, in mid-2026, also the most useful, and the buyer who sits in it is buying real measured throughput at the lowest cost per measured throughput dollar in the category. There is no second part to the argument.
