VoltgridNews

ASUS ProArt GR1X Review: 40 tok/s on a 27B Model, $4,499 for 64 GB

The $4,499, 64 GB ASUS ProArt GR1X hits 40.0 tok/s on a 27B model in StorageReview's llama.cpp tests, ahead of RTX Spark laptops; no 70B or 120B run has been published.

Local AI Desk

10 min read

Follows ASUS ProArt GR1X and MSI EdgeMesa N Hit 40 tok/s in StorageReview Tests

Silver ASUS ProArt GR1X mini PC with a perforated front grille and NVIDIA RTX Spark badge on a black background

The ASUS ProArt GR1X, as shown on ASUS's product page. The 64 GB model is the one StorageReview tested.

Image: ASUS

Source · Press image from the ASUS product page, used for news reporting.

The ASUS ProArt GR1X served 40.0 tokens per second on Qwen3.8 27B (Q4_K_M, llama.cpp CUDA) with eight requests in StorageReview's Nov. 18 review, one of three reviews published that day that this piece draws on. ASUS launched the ProArt GR1X in the US on Nov. 18, 2026, listing the 64 GB model at $4,499, though StorageReview found the ASUS eShop out of stock and Best Buy showing coming soon.

That figure sets the 64 GB GR1X level with MSI's 128 GB EdgeMesa N on the same NVIDIA RTX Spark N1X chip, which StorageReview measured at 39.8 tokens per second in the identical test, as Voltgrid reported on Nov. 18. This review adds Windows Central's UL Procyon results on 64 GB and 128 GB units, Notebookcheck's first look at a 128 GB unit, ASUS's spec page and the prices, and sets the GR1X against its closest rivals.

The short version for a buyer: at $4,499 the 64 GB GR1X sits $100 above the $4,399 entry price NVIDIA quotes for RTX Spark desktops, and $100 above MSI's 64 GB EdgeMesa N, which Notebookcheck reported cut to $4,399 on Nov. 18. In StorageReview's llama.cpp tests it matches MSI's dearer 128 GB box on 8B to 27B models, and it is 18% to 25% ahead of two RTX Spark laptops at eight requests; Windows Central's Procyon scores put it 10.6% to 13.1% ahead, and Notebookcheck found CPU performance essentially identical to laptops. What nobody has measured yet is the reason many people would buy a 128 GB box: a 70B or 120B model on this machine.

What's new

The GR1X is ASUS's first desktop on the RTX Spark N1X, the Arm chip that pairs a 20-core NVIDIA Grace CPU with a 6,144-core Blackwell GPU and that ASUS already ships in its ProArt P16 and P14 laptops. The change is power. Notebookcheck reports that the N1X is held to 80 W in laptops, while the desktop chip may draw up to 140 W; StorageReview says ASUS built the dual-fan cooler to sustain 140 W, and Windows Central gives the same 140 W ceiling. Notebookcheck adds that the 140 W figure applies mainly when the CPU and GPU are loaded together, with the GPU drawing up to 95 W.

Memory is the second change. ASUS's spec page, read on Nov. 20, lists 64 GB or up to 128 GB of LPDDR5X unified memory, shared by the CPU and GPU, at 300 GB/s. Notebookcheck's 128 GB test unit carries LPDDR5X-9400. The 128 GB option matches the top RTX Spark laptops, but in a 150 x 150 x 51 mm box that weighs 1.48 kg, according to ASUS.

Storage is the third. ASUS lists two M.2 slots, one PCIe 5.0 and one PCIe 4.0, for up to 4 TB in all, and StorageReview calls the second slot, which takes standard retail drives for models and datasets, the design's main win. The rest of the spec sheet is a 10G RJ45 port, Wi-Fi 7, four USB 3.2 Gen 2x2 Type-C ports (three with DisplayPort output, one for power input), one HDMI 2.1 port and a 240 W USB-C adapter.

The software is Windows 11 on Arm, Home or Pro. ASUS and NVIDIA credit the platform with up to 1 petaFLOP of FP4 AI compute, a figure published without a dense or sparse label.

What it solves

The GR1X puts the RTX Spark platform on a desk near the $4,399 entry price NVIDIA quotes. NVIDIA told StorageReview that RTX Spark desktops start at $4,399, which puts the 64 GB GR1X $100 above the entry price, and Notebookcheck reported on Nov. 18 that MSI's 64 GB EdgeMesa N now lists at that $4,399. For a team that wants a Windows box running local models without a cloud account, that is the price of admission.

ASUS says the machine runs language models of up to 120 billion parameters with up to 1 million tokens of context, and it ties the 70B and 120B class to the 128 GB configuration. That is a maker's claim, which StorageReview notes ASUS frames as capacity guidance that assumes suitable quantization. No outlet has yet published a 70B or 120B measurement on any RTX Spark system, so whether a 120B model runs at a usable speed on a 300 GB/s memory bus is still unknown.

What the 64 GB model does solve, on the evidence so far, is serving small and mid-size models to several users at once. In StorageReview's llama.cpp tests (CUDA build, Q4_K_M quantization, Windows 11 Pro, its sr-client 1.13 harness), the 64 GB unit generated 13.1 tokens per second on Qwen3.8 27B with one request, 31.4 with four and 40.0 with eight, with a time to first token of 8.6 seconds. Ministral 3 3B ran at 58.4 tokens per second and Ministral 3 8B at 42.7 with one request; the 8B model reached 109.3 tokens per second across four requests.

The limit is memory. StorageReview did not run its 35B coding-agent test or the Qwen-Image Q8 image model on the 64 GB unit, because they need the 128 GB tier, and its gpt-oss:20b coding-agent run passed only 1 of 3 tasks at 57.1 to 58.5 tokens per second, which the outlet put down to run-to-run variance.

Pros and Cons

The items above rest on these measurements. StorageReview measured the 27B serving results in llama.cpp with Q4_K_M weights and found the single-request results of all four RTX Spark systems within 6% of each other, while the eight-request result put the desktops ahead. In MLPerf Client v2.0, StorageReview measured 56.8 tokens per second on Llama 3.1 8B and 96.3 on Phi-4-mini, a four-way tie with the EdgeMesa N and two laptops. StorageReview also found the GR1X ran 3 to 8 C cooler than the EdgeMesa N in its CPU tests. Windows Central measured no more than 55 dB in a stress test and says the box does not throttle; Notebookcheck calls it very quiet in early tests and has no noise figures yet. On price, the 64 GB GR1X at $4,499 is $3,000 below the $7,499 that Notebookcheck and HotHardware reported for MSI's 128 GB, 4 TB EdgeMesa N on Nov. 18.

On the downside, StorageReview's unit shipped with a 1 TB PCIe 4.0 SSD that read at 6,511 MB/s; a cold load of gpt-oss:20b, a 12.8 GB model, took 10.94 seconds to the first token against 7.00 seconds on the EdgeMesa N and 6.70 seconds on HP's laptop, both with PCIe 5.0 drives. Windows Central lists weak port selection and placement and a "super expensive" price among its reasons to avoid the machine. Notebookcheck reports that some apps and peripherals still do not work on Windows on Arm, and that Call of Duty, Battlefield and F1 2025 do not run, while Adobe's apps do.

Who it's for

The GR1X fits a developer or small team that serves 8B to 27B models to a handful of users from one Windows machine. StorageReview's 40.0 tokens per second across eight requests on a 27B model at Q4_K_M, and 109.3 across four on an 8B model, are the numbers that matter here, and the 64 GB, $4,499 configuration delivers them.

It also fits a creative studio that already works in Adobe's apps and wants local image generation beside them. Windows Central measured 20.55 seconds per SDXL image on the 64 GB unit in UL Procyon, and 20.6 seconds on the 128 GB unit, against 25.05 seconds on a 128 GB Surface Laptop Ultra.

A buyer who wants a 70B or 120B model should wait for the 128 GB configuration to be priced and measured. ASUS ties that class to 128 GB, has not published a US price for it, and no outlet has run such a model on the GR1X.

It does not fit a gamer or anyone whose tools depend on x86 drivers: Notebookcheck's list of games that do not run on Windows on Arm, and its warning about peripherals, are the reason.

How it compares

The closest rival is MSI's EdgeMesa N, reviewed here, the other RTX Spark mini PC StorageReview tested. On StorageReview's 27B llama.cpp test at eight requests it measured 39.8 tokens per second to the GR1X's 40.0, with 13.0 against 13.1 at one request: the same speed. MSI cut the 128 GB, 4 TB EdgeMesa N to $7,499, according to Notebookcheck and HotHardware on Nov. 18, which is also the MSRP NVIDIA gave StorageReview; MSI's own US store listed it at $7,999 when read on Nov. 11, out of stock, as Voltgrid reported. Against $7,499, the 64 GB GR1X costs $3,000 less, with half the memory and a quarter of the storage. A 64 GB EdgeMesa N with 1 TB now lists at $4,399, $100 below the GR1X, according to Notebookcheck. MSI's page lists 273 GB/s of memory bandwidth for the same N1X chip; ASUS lists 300 GB/s. Neither maker has explained the gap, and the two boxes measured the same speed in StorageReview's tests.

StorageReview measured 40.0 and 39.8 tok/s on the same 27B test; the 300 and 273 GB/s bandwidth figures are the makers' listings.

Graphic: Voltgrid News

Against the RTX Spark laptops, which Voltgrid reviewed, the desktop gains most under load. In StorageReview's 27B test at eight requests, HP's laptop reached 33.8 tokens per second and Dell's XPS 16 31.9, so the GR1X was 18% and 25% ahead. Windows Central measured with a different tool, UL Procyon's text generation suite, so its numbers do not mix with StorageReview's: it recorded 74.9 tokens per second on Phi-3.5-mini for the 64 GB GR1X, 77.95 for the 128 GB GR1X and 72.21 for a 128 GB Surface Laptop Ultra, and 28.1, 28.8 and 27.44 on a 13B Llama model. Windows Central puts the desktop's lead at 10.6% to 13.1% on Procyon's scores, which are not the same as the tokens-per-second gaps.

Corsair's FlexPrime V20C, an Intel Arc Pro B70 tower at $3,999.99 with 32 GB of GDDR6 graphics memory at 608 GB/s, ran the same Qwen3.8 27B Q4_K_M model in StorageReview's llama.cpp server test at 19.8 tokens per second with one request and 35.4 with eight, as Voltgrid reported. The model and quantization match, but the harness, GPU backend and operating system differ: the Corsair ran llama.cpp server on Ubuntu, the GR1X StorageReview's Windows sr-client suite. That makes the result indicative, not a like-for-like comparison. The Corsair also cannot hold more than 32 GB of model on its GPU.

For larger models, the ASUS Ascent GX10 is ASUS's other answer: a 128 GB NVIDIA GB10 box on Linux at $5,999, listed on Oct. 30, and out of stock with notify-me only on ASUS's US store on Nov. 11. StorageReview measured it on gpt-oss-120b in vLLM at 70 tokens per second with one request and 680 in aggregate at batch 64. That is a different model, engine and operating system from the GR1X tests, so the two numbers are not comparable; what they show is that a 120B result exists for the GX10 and not yet for the GR1X.

Why Memory Bandwidth Sets Local Model Speed

A language model generates one token at a time and re-reads its weights from memory for each one, so on a machine like the GR1X speed depends on memory bandwidth as much as on compute. The N1X shares one pool of LPDDR5X between CPU and GPU: ASUS's spec page lists 300 GB/s, and ServeTheHome reports LPDDR5X-9400 on a 256-bit bus, up to 300 GB/s.

MSI lists 273 GB/s for the same chip. StorageReview, though, reports LPDDR5X-9400 on all four of its RTX Spark systems and measured STREAM Triad at 124,529 MB/s on the GR1X and 122,423 MB/s on the EdgeMesa N. Those are one lab's measurements, not a ruling on MSI's figure. Capacity sets the other limit: ASUS ties 70B and 120B models to the 128 GB tier, and no outlet has timed one.

What to watch

Notebookcheck said on Nov. 18 that its full review, with noise and power measurements, would follow within days. ASUS has not published a US price for the 128 GB GR1X, and the 64 GB model was out of stock on ASUS's US store and listed as coming soon at Best Buy on Nov. 18, according to StorageReview. The first 70B or 120B measurement on a 128 GB RTX Spark machine will decide whether ASUS's 120B claim holds at a usable speed. And the 273 GB/s and 300 GB/s figures for the same chip remain unreconciled until NVIDIA, ASUS or MSI explains them.

How it compares

Full table
ASUS ProArt GR1X Review: 40 tok/s on a 27B Model, $4,499 for 64 GB: how the machine compares with other local AI hardware
ProductMemoryBandwidthAI computeMeasured speedPrice
NVIDIA RTX PRO 6000 Blackwell Workstation EditionNVIDIA (sold through PNY, TD SYNNEX and workstation makers)96 GB[16]1,792 GB/s[16]4,000 AI TOPS (FP4, with sparsity) (sparse)[16]gpt-oss-120b MXFP4, stock vLLM, 512/512 chat workload: 3,064 output tok/s at 128 streams (1,169 on 8,192/1,024 prompts); Server Edition at 600W (StorageReview)[17]$16,000 on NVIDIA's US marketplace (Workstation Edition), listed out of stock[18]
Apple Mac Studio (M5 Ultra)Appleup to 512 GB unified (96 GB on 30-core M5 Ultra; 256 GB or 512 GB on 36-core/80-core M5 Ultra)[19]1.2 TB/s[19]—GLM-5.3-Flash MLX 4-bit with MTP on TensorFold, ISL 2048/OSL 512: 94.0 tok/s at 0 context, 96.5 at 128K, 79.4 at 512K; TTFT 1.1 s at 0, 66.3 s at 128K (Tom's Hardware, 256 GB M5 Ultra)[20]$22,299 minimum with 512 GB ($6,799 36-core CPU/80-core GPU M5 Ultra config plus $15,500 memory upgrade; Oct 28, 2026)[21]
Corsair FlexPrime V20C Workstation (CSW-9040004-NA)Corsair32 GB GDDR6 on the Arc Pro B70 GPU + 32 GB (2x 16 GB) DDR5 system memory[22]608 GB/s (GPU GDDR6)[22]—Qwen3.8 27B Q4_K_M, llama.cpp server: 19.8 tok/s at 1 request, 33.5 at 4, 35.4 at 8; TTFT 9,005 ms on 8,192-token prompt (StorageReview)[23]$3,999.99 on Corsair US store (32 GB DDR5, 1 TB SSD; read Nov 15, 2026)[22]
ASUS ProArt GR1XASUSThis story64 GB or up to 128 GB LPDDR5X unified memory (ASUS spec page, read Nov 18, 2026)[24]300 GB/s (ASUS spec page, read Nov 18, 2026)[24]WITHDRAWN, not a sparse figure: ASUS and NVIDIA say 'up to 1 petaFLOP FP4' with no sparse or dense label; no field asserts either[25]Qwen3.8 27B Q4_K_M, llama.cpp CUDA, Windows 11: 13.1 tok/s output at 1 request, 31.4 at 4, 40.0 at 8 requests; TTFT 8.6 s (StorageReview, 64 GB unit, sr-client 1.13)[26]$4,499 (64 GB, 1 TB Gen4, Windows 11 Home) at ASUS eShop US; $4,499.99 at Best Buy, which lists 64 GB/2 TB at $4,999.99 (StorageReview, Nov 18, 2026)[26]
Mechrevo Master Road X10Mechrevo64 GB (with 1 TB SSD) or 128 GB (with 2 TB SSD) unified memory[27]300 GB/s (unified memory)[27]up to 1 PFLOP FP4 (sparse)[27]—CNY 32,999 (64 GB/1 TB) to CNY 49,999 (128 GB/2 TB), China only[27]
NVIDIA RTX Spark (N1X) PCsNVIDIA (MediaTek collaborated on the custom CPU design)Up to 128 GB LPDDR5X (20-core CPU / 6,144-core GPU version); up to 64 GB (18-core / 5,120-core laptop version)[25]Up to 300 GB/s: LPDDR5X-9400 on a 256-bit bus (NVIDIA's product page lists no bandwidth)[28]—gpt-oss:20b via Ollama in a Codex CLI coding agent, 64K context: 58.5 tok/s generation, 3 of 4 tasks passed (Dell XPS 16, 20-core N1X, 64 GB; StorageReview)[29]Laptops from $2,599; pricing and configurations vary by region[30]
AMD Ryzen AI Halo (Ryzen AI Max+ PRO 495, 192 GB)AMD192 GB LPDDR5X (256-bit)[31]273 GB/s[31]——$7,299.99 for the Linux version (released the week of Oct 12, 2026); Windows version coming soon[31]
ASUS Ascent GX10ASUS64 GB or 128 GB LPDDR5X-8533 unified[32]273 GB/s (8533 MT/s), as listed by StorageReview; ASUS's spec page states no bandwidth[33]Up to 1 PFLOP FP4 (GB10; NVIDIA's footnote: theoretical, using the sparsity feature). ASUS's pages say 1 petaFLOP / FP4 with no sparsity label (sparse)[34]gpt-oss-120b, vLLM online serving, equal ISL/OSL: 70 tok/s at batch 1 rising to 680 tok/s aggregate at batch 64 (StorageReview)[33]$5,999 (128 GB, 1 TB, GX10-GG0010BN, ASUS eShop US, Oct 30, 2026)[35]
Show the other 34 machines
ASUS ProArt GR1X Review: 40 tok/s on a 27B Model, $4,499 for 64 GB: how the machine compares with other local AI hardware: the other machines
ProductMemoryBandwidthAI computeMeasured speedPrice
ASUS ExpertCenter Pro ET900N G3ASUS (NVIDIA DGX Station GB300 platform)748 GB coherent: 252 GB HBM3e (GPU) + 496 GB LPDDR5X (CPU)[1]7.1 TB/s (GPU HBM3e); 396 GB/s (CPU LPDDR5X)[1]up to 20 PFLOPS AI performance (ASUS gives no precision or sparsity label) (sparse)[2]GPT-OSS-120B MXFP4, vLLM, 512/512, 1-128 streams: 9,280 aggregate output tok/s at 128 streams; 5,023 on 8,192/1,024 (StorageReview, ASUS-supplied unit, no add-in GPU)[3]$111,999.99 at Central Computers, US (ET900N G3-2SSD-V2, read Nov 8, 2026); no ASUS list price[4]
HP ZGX Fury AI StationHP252 GB HBM3e GPU + 496 GB LPDDR5x CPU (4x 128 GB SOCAMM); 748 GB unified[5]7.1 TB/s (GPU HBM3e); 396 GB/s (CPU LPDDR5x)[5]up to 20 petaFLOPS FP4 (HP gives no sparsity label) (sparse)[5]Two-unit cluster with an ASUS ET900N G3 over two 400G DACs: GLM-5.3 NVFP4 5,018 output tok/s at 256 streams, 512/512, vLLM 0.27.1 TP2 with MTP (StorageReview); no single-unit HP result[6]$132,000.00 on HP Store US (ZGX Fury G1n, 3-yr warranty, no graphics card, 4x 2 TB SED SSDs; Add to cart; read Nov 11, 2026)[7]
MSI XpertStation WS300MSI (NVIDIA DGX Station platform)252 GB HBM3e GPU + 496 GB LPDDR5X CPU (748 GB coherent)[8]7.1 TB/s (GPU HBM3e); 396 GB/s CPU LPDDR5X[8]—StorageReview vLLM: gpt-oss-120b MXFP4 9,294 output tok/s at 128 streams (512/512), 5,038 on 8,192/1,024; MiniMax M3 NVFP4 1,041 at 32 streams with EAGLE3-GQA at a forced acceptance length of 3 (best-case)[8]$99,999.00 'Start From' at Newegg US (model WS300T60L, 2x 2 TB NVMe RAID1), request-a-quote listing sold by Newegg (read Nov 15, 2026)[9]
NVIDIA DGX StationNVIDIA (systems built by ASUS, Dell, GIGABYTE, HP, MSI, Supermicro and Exxact)748 GB coherent: 252 GB HBM3e (GPU) + 496 GB LPDDR5X (CPU)[10]GPU HBM3e 7.1 TB/s; CPU LPDDR5X 396 GB/s; NVLink-C2C 900 GB/s between them[10]15 PFLOPS FP4 (without sparsity)[10]gpt-oss-120b MXFP4, vLLM, 512/512 chat workload: 9,280 output tok/s at 128 streams (MSI WS300 9,294), single GB300 (StorageReview, ASUS ExpertCenter Pro ET900N G3)[3]No NVIDIA list price; lowest US partner listing read Nov 4, 2026: Exxact Valence VWS-158270643 from $98,450 (Tom's Hardware found $94,930 in Aug); Dell Pro Precision with GB300 lists $160,195[11]
Supermicro Super AI Station (ARS-511GD-NB-LCC)Supermicro (NVIDIA DGX Station platform)748 GB total coherent: 252 GB HBM3e GPU + 496 GB LPDDR5X system[12]7.1 TB/s (GPU HBM3E); 396 GB/s (CPU LPDDR5X)[13]up to 20 PFLOPS FP4 (Supermicro gives no sparsity label; ServeTheHome describes it as sparse FP4)[12]Qwen3.8 Flash Next NVFP4, vLLM, tool-loop 4,096-in/512-out: several scenarios exceeded 3,000 output tok/s at 32 concurrent users, still rising (ServeTheHome; RTX PRO 2000 installed)[14]$92,887.47 before tax and shipping on Supermicro's online store (ARS-511GD-NB-LCC-01-G2); about $90,101.05 with a 3% new-customer discount (TechRadar, Sep 1, 2026)[15]
Dell Pro Precision with GB10 (model FCM1253; launched as Dell Pro Max with GB10)Dell128 GB LPDDR5x up to 8533 MT/s onboard (64 GB option also listed)[36]273 GB/s (128 GB LPDDR5x, 256-bit)[37]—gpt-oss-120b, vLLM online serving, equal ISL/OSL: 68.57 tok/s at batch 1 rising to 670.83 tok/s aggregate at batch 64 (StorageReview)[38]$7,901.00 (default FCM1253, 128 GB, 2 TB Gen4 QLC) on Dell.com US, unchanged; 64 GB memory option listed at -$2,150.00, 4 TB TLC SED SSD +$2,450.00 (read Nov 20, 2026)[39]
Framework Desktop (Ryzen AI Max+ PRO 495, 192 GB)Framework192 GB LPDDR5X-8533 soldered[40]273 GB/s (256-bit)[40]——$6,799 (Max+ PRO 495 192 GB DIY Edition pre-order, Nov 3, 2026; storage not included)[41]
GIGABYTE AI TOP ATOMGIGABYTE128 GB unified system memory[42]273 GB/s (LPDDR5x-8533, 256-bit)[43]—gpt-oss-120b, vLLM online serving, equal ISL/OSL: 60.71 tok/s at batch 1 rising to 649.62 tok/s aggregate at batch 64 (StorageReview)[43]$7,499.99 (128 GB, 4 TB PCIe 4.0) and $9,405.63 (128 GB, 4 TB PCIe 5.0; more options from $7,999.99); 128 GB from $5,999.99 (1 TB) and $6,999.99 (2 TB) at Newegg US, read Nov 9, 2026[44]
GMKtec EVO-X5 Pro AI Mini PCGMKtec192 GB LPDDR5X-8533 onboard (8 x 24 GB, non-upgradeable); up to 160 GB allocatable as VRAM[45]Up to 273 GB/s[46]—GMKtec Laboratory table: GLM-5.3-Flash UD-IQ4_XS (146.05 GB) decode 17.79 tok/s, prefill 70.18, TTFT 14.71 s; Qwen3.8-Flash-Next ROCmFP4-FAST (93.48 GB) decode 45.08 tok/s[45]$6,799 (192 GB/2 TB) and $7,099 (192 GB/4 TB), all variants available, GMKtec US store, read Nov 9, 2026[45]
HP ZGX Nano AI Station (G1n)HP128 GB coherent unified system memory[47]273 GB/s (128 GB LPDDR5x)[48]GB10: up to 1 PFLOP FP4, theoretical, using the sparsity feature; HP and reviewers list 1,000 TOPS FP4 (sparse)[34]gpt-oss-120b, vLLM online serving, equal ISL/OSL: 69.6 tok/s at batch 1 rising to 722.9 tok/s aggregate at batch 64 (StorageReview)[49]$7,399.00 at the HP Store for the 4 TB model; $4,759 on Amazon (Notebookcheck, review updated Oct 5, 2026)[50]
Lenovo ThinkStation PGXLenovo128 GB LPDDR5X unified (256-bit)[51]273 GB/s[51]Up to 1 PFLOP FP4 (Lenovo PSREF gives no sparsity label; XDA says NVIDIA's figure assumes structured sparsity) (sparse)[51]Qwen3-Coder-Next FP8 (FP8 KV cache), vLLM NGC container, one request at a time, 170K context: 25-40 tok/s (XDA Developers, Lenovo-loaned unit)[52]US from $5,079 (1 TB base model) on Lenovo's site, per TechRadar's review (Jun 9, 2026); UK from GBP 5,200[53]
MSI EdgeMesa N AI+ (128 GB, X-005US)MSI128 GB LPDDR5X unified system memory[54]273 GB/s[54]—Qwen3.8 27B Q4_K_M, llama.cpp CUDA, Windows 11: 13.0 tok/s output at 1 request, 31.1 at 4, 39.8 at 8 requests (StorageReview, 128 GB unit, default 62.5 GB GPU allocation)[55]$7,499.00 (128 GB, 4 TB, X-005US, MSI US store, out of stock, read Nov 21, 2026)[54]
MSI EdgeXpert (MS-C931)MSI64 GB or 128 GB LPDDR5x unified (256-bit)[56]273 GB/s[56]1 PFLOPS FP4 (sparse)[56]—No list price; MSI sells via quote (Get a quote)[56]
NVIDIA DGX SparkNVIDIA (partner GB10 systems from Acer, ASUS, Dell, GIGABYTE, HP, Lenovo and MSI)128 GB; a 64 GB configuration is sold only through OEM partners[34]273 GB/s[34]Up to 1 PFLOP FP4 (theoretical, using the sparsity feature) (sparse)[34]Qwen 3.8 27B (dense) with multi-token prediction: decode of about 20 tokens/s, with time-to-first-token staying interactive out to the full 262K context.[57]$6,950.00 (128 GB, 4 TB) on NVIDIA's enterprise marketplace with an Add To Cart button in a copy cached 07:59 UTC on 2026-10-31; stock unverified (listed out of stock on 2026-10-21); 64 GB partner systems from $4,999[58]
NVIDIA DGX Spark 64 GBNVIDIA (sold by Acer, ASUS, Dell, Gigabyte, HP and MSI)64 GB LPDDR5X unified (two units pool 128 GB over ConnectX-7)[59]273 GB/s (256-bit; NVIDIA spec table covers both 64 GB and 128 GB)[34]1 PFLOPS FP4 (sparse; NVIDIA spec table covers both memory configurations)[34]—from $4,999 (partner systems, Oct 2026; no Founders Edition)[59]
Sipeed MaixStationSipeed128 GB LPDDR5X (256-bit)[31]273 GB/s (Sipeed claim)[31]256 TOPS FP4 (native FP4, no sparsity; Axera/Sipeed claim)[31]—about $3,999 expected for 128 GB, with Kickstarter launch discounts (Sipeed to Tom's Hardware, Oct 2026)[31]
AMD Ryzen AI Halo Developer PlatformAMD128 GB LPDDR5x-8000 unified[60]256 GB/s[60]—gpt-oss-120b, vLLM online serving, equal ISL/OSL 256/256: 222 tok/s aggregate at batch 64 vs 701 on DGX Spark (StorageReview)[61]$3,999 with 2 TB SSD (Jul 2026)[61]
Framework Desktop (Ryzen AI Max+ 395)Framework128 GB LPDDR5x-8000 soldered (32 GB and 64 GB options)[62]256 GB/s[63]—gpt-oss-120b MXFP4 in LM Studio (out-of-the-box setup): roughly 45 tok/s (ServeTheHome; Ryzen AI Max+ 395, 128 GB, Windows 11 Pro)[64]$3,449 (DIY Edition, Max+ 395 with 128 GB, system only; out of stock, Nov 3, 2026)[65]
Acer Veriton GN100 AI Mini WorkstationAcer128 GB LPDDR5x coherent unified[66]——gpt-oss-120b, vLLM, total token throughput: prefill heavy 303.61 to 2,777.56 tok/s; decode heavy 38.38 to 292.41 tok/s (batch 1 to 64, StorageReview)[67]from $3,999 in North America (announced Sep 3, 2025; prices vary by region)[66]
ASRock Industrial AI BOX-A395ASRock Industrial128 GB LPDDR5X-8000 soldered[68]——AgentSTH V7, 32 threads: 15% ahead of the next-closest Ryzen AI Max+ 395 box (single-core within 1%); estimated SPEC CPU2026 int rate: top of a 44% spread across four 395 systems; measured by ServeTheHome, Ubuntu 26.04[69]—
Beelink GTR9 Pro (Ryzen AI Max+ 495, 128 GB)Beelink128 GB LPDDR5X-8533 unified (soldered)[70]—132 AI TOPS on Beelink's page; AMD lists the PRO 495 at Overall TOPS up to 132 and NPU TOPS up to 55; no precision stated by either[70]4.24 tok/s, LM Studio, Qwen3.8-35B-A3B (Beelink's own figure on its store page; quantization, context, backend and RAM size not stated)[70]$4,799 (128 GB/2 TB), Beelink store, pre-order (read Nov 13, 2026)[70]
Beelink GTR9 Pro (Ryzen AI Max+ 495, 192 GB)Beelink192 GB LPDDR5X-8533 unified (soldered)[70]—132 AI TOPS on Beelink's page; AMD lists the PRO 495 at Overall TOPS up to 132 and NPU TOPS up to 55; no precision stated by either[70]4.24 tok/s, LM Studio, Qwen3.8-35B-A3B (Beelink's own figure on its store page; quantization, context, backend and RAM size not stated)[70]$7,399 (192 GB/2 TB) or $7,799 (192 GB/4 TB), Beelink store, pre-order (read Nov 13, 2026)[70]
Bosgame M5 MAX AI Mini PCBosgame192 GB unified LPDDR5 (soldered); up to 160 GB allocatable to the GPU[71]———$6,199 introductory first-batch pre-order price (192 GB, 2 TB PCIe 4.0 SSD), from Nov 3, 2026, until the first allocation sells out[71]
Dell Pro Max Micro (FCM2250)Dell24 GB GDDR7 on the RTX PRO 4000 SFF (23.9 GB usable measured) + 96 GB DDR5-6400 CSODIMM system RAM in StorageReview's tested build[72]——Ministral 3 8B, llama.cpp server (CUDA), 4,096-in/1,024-out: 52.0 tok/s at 1 request, 108.9 at 4; TTFT 1,477 ms (StorageReview, RTX PRO 4000 SFF 24 GB, driver 595.99.02, Ubuntu 26.04.1)[72]From $1,449.00 on Dell.com US; RTX PRO 4000 SFF 24 GB option +$3,870.00 (read Nov 16, 2026)[73]
Dell Pro Precision 9 T4DellUp to 2x 96 GB GDDR7 (two RTX PRO 6000 Max-Q) + up to 4 TB DDR5 RDIMM in 16 slots, 8 channels[74]——MiniMax M2.7 Q4_K_M (129.3 GB) split across 2x RTX PRO 6000 Max-Q: 83.7 tok/s one user, 127.3 peak; 30.3 on one card plus system RAM (StorageReview, llama.cpp b11716, Ubuntu 26.04.1)[74]From $5,925.00 on Dell.com US; dual RTX PRO 6000 Max-Q option +$30,820.00 (read Nov 14, 2026). StorageReview's tested build (Xeon 698X, 256 GB, 2x Max-Q) about $66,600[75]
Dell Pro Precision with GB300 (model FCT6263; announced as Dell Pro Max with GB300)Dell252 GB HBM3e GPU + 496 GB LPDDR5X CPU[76]———$160,195.00 (Dell.com US default config FCT6263, read Nov 8, 2026)[76]
Gigabyte AI TOP Aeroup to 128 GB LPDDR5X unified[77]————
GMKtec EVO-X2 AI Mini PCGMKtec64 GB or 128 GB onboard LPDDR5X-8000 (non-upgradeable)[78]——Llama 3.3 70B Q6_K Instruct: 3.7-3.8 tok/s generation fully on the Radeon 8060S (about 2 tok/s CPU-only); Qwen3-32B 10.4-10.5 tok/s; 128 GB unit, short prompt, tool not named (ServeTheHome)[79]$3,499.99 (128 GB/1 TB) and $3,649.99 (128 GB/2 TB), each against a $3,999.99 compare-at price (GMKtec US store, read Nov 11, 2026)[78]
HP Z2 Mini G1a WorkstationHPup to 128 GB unified, up to 96 GB assignable to the GPU, up to 8000 MT/s[80]—up to 50 NPU TOPS[80]gpt-oss-120b GGUF (63.39 GB, full GPU offload) in LM Studio: close to 40 tok/s (StorageReview)[81]$4,781 list for StorageReview's review configuration; about $3,300 street (B&H top spec with 2 TB SSD at $3,342.65, Aug 2025)[81]
Microsoft Surface RTX Spark Dev BoxMicrosoft128 GB unified memory[82]—up to 1 PFLOP of AI compute (Microsoft; precision and sparsity not stated on the page) (sparse)[82]—From $5,999.99 (Microsoft Store US, pre-order, read Nov 1, 2026)[82]
MSI EdgeMesa N AI+ (64 GB)————$4,399 (64 GB / 1 TB, after MSI's cut from $4,699; Nov 2026)[83]
Tenstorrent TT-QuietBox 2 (Blackhole)Tenstorrent128 GB GDDR6 on the accelerators + 256 GB DDR5-5600 system memory[84]———$9,999 (Oct 30, 2026)[84]
WiCi One (RTX 5060 Ti edition)WiCi16 GB GDDR7 VRAM; 1 TB NVMe 4 SSD for TurboStream weight paging (WiCi's claim: NVMe used as working VRAM). System RAM not stated[85]———$2,599 list (WiCi shop, Oct 2026); a $600-off code (WICIWELCOME) is applied at checkout. Early-access price was $1,999 (Notebookcheck, Engadget, Sep 2026)[85]
WiCi One (RTX 5090 edition)WiCi32 GB GDDR7 VRAM; 4 TB NVMe 5 SSD for TurboStream weight paging (WiCi's claim: NVMe used as working VRAM). System RAM not stated[85]———$9,999 list (WiCi shop, Oct 2026); a $600-off code (WICIWELCOME) is applied at checkout[85]

Italic values are reported by an outlet or measured by an independent lab; the rest come from the maker. Measured speeds use different models, quantizations and tools, so read each with its source rather than against the others. Voltgrid News does not test hardware itself.

Sources (85)
  1. [1] ASUS
  2. [2] ASUS
  3. [3] StorageReview
  4. [4] Central Computers
  5. [5] HP
  6. [6] StorageReview
  7. [7] HP
  8. [8] StorageReview
  9. [9] Newegg
  10. [10] NVIDIA
  11. [11] Exxact
  12. [12] Supermicro
  13. [13] ServeTheHome
  14. [14] ServeTheHome
  15. [15] TechRadar
  16. [16] NVIDIA
  17. [17] StorageReview
  18. [18] NVIDIA
  19. [19] Apple
  20. [20] Tom's Hardware
  21. [21] MacRumors
  22. [22] Corsair
  23. [23] StorageReview
  24. [24] ASUS
  25. [25] NVIDIA
  26. [26] StorageReview
  27. [27] Notebookcheck
  28. [28] ServeTheHome
  29. [29] StorageReview
  30. [30] NVIDIA
  31. [31] Tom's Hardware
  32. [32] ASUS
  33. [33] StorageReview
  34. [34] NVIDIA
  35. [35] ASUS eShop
  36. [36] Dell
  37. [37] HotHardware
  38. [38] StorageReview
  39. [39] Dell
  40. [40] Framework Blog
  41. [41] Framework
  42. [42] GIGABYTE (PR Newswire)
  43. [43] StorageReview
  44. [44] Newegg
  45. [45] GMKtec
  46. [46] GMKtec
  47. [47] HP
  48. [48] HotHardware
  49. [49] StorageReview
  50. [50] Notebookcheck
  51. [51] Lenovo PSREF
  52. [52] XDA Developers
  53. [53] TechRadar
  54. [54] MSI
  55. [55] StorageReview
  56. [56] MSI
  57. [57] Tom's Hardware
  58. [58] NVIDIA Marketplace
  59. [59] NVIDIA Blog
  60. [60] AMD
  61. [61] StorageReview
  62. [62] Framework
  63. [63] Framework Blog
  64. [64] ServeTheHome
  65. [65] Framework
  66. [66] Acer
  67. [67] StorageReview
  68. [68] ServeTheHome
  69. [69] ServeTheHome
  70. [70] Beelink
  71. [71] Notebookcheck
  72. [72] StorageReview
  73. [73] Dell
  74. [74] StorageReview
  75. [75] Dell
  76. [76] Dell
  77. [77] Notebookcheck
  78. [78] GMKtec
  79. [79] ServeTheHome
  80. [80] HP
  81. [81] StorageReview
  82. [82] Microsoft
  83. [83] Notebookcheck
  84. [84] Tenstorrent
  85. [85] WiCi

Sources

Published
Sources
4 cited