VoltgridNews

MSI EdgeMesa N Review: 39.8 tok/s on a 27B Model, 35B Agent, Now $7,499

StorageReview and HotHardware measured MSI's 128 GB RTX Spark mini PC: it ties ASUS's GR1X on a 27B model, passed a 35B coding agent 3 of 3 times and stays near silent. MSI cut its price to $7,499.

Local AI Desk

Updated 10 min read

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

MSI EdgeMesa N AI+ RTX Spark mini PC, front view of the aluminium box with MSI badge and vented grille

The MSI EdgeMesa N AI+, as shown in NVIDIA's RTX Spark launch post.

Image: NVIDIA / MSI

Source · Press image from the NVIDIA blog, used for news reporting.

Update (Nov 20, 19:14 UTC): Tom's Hardware published its review of the 128 GB EdgeMesa N AI+ on Nov 20. Under its AgentPerf-Local LLM test it measured about 140 W at the wall and 38.2 dBA at 18 inches (46 cm), and about 200 W and 40.7 dBA under gaming loads. On Qwen 3.8 27B with DFlash2 speculative decoding, the EdgeMesa performed practically the same as Microsoft's Surface Laptop Ultra, while AMD's Ryzen AI Halo, on Windows with the Vulkan backend, took about four times as long to finish. Tom's listed the price, limited ports and the lack of DGX Spark-style clustering as cons.

The 128 GB MSI EdgeMesa N AI+ served Qwen3.8 27B at 39.8 tokens per second across eight simultaneous requests in StorageReview's November 18, 2026 review, and it was the only RTX Spark mini PC in that lab to finish a 35B coding-agent task on all three attempts.

HotHardware, also reviewing the 128 GB unit on November 18, found the 1.2-litre box did not register on its sound meter and drew about 205 W at the wall at its peak in a gaming load. The same day, Notebookcheck reported that MSI had cut the 128 GB, 4 TB model to $7,499, $500 below the $7,999 MSI's US store listed when read on November 11; $7,499 is also the MSRP NVIDIA gave StorageReview and the price HotHardware gave. As Voltgrid reported on November 18, StorageReview's serving numbers put the EdgeMesa level with ASUS's ProArt GR1X; this review sets those numbers against the rest of the evidence and the price.

What's new

The EdgeMesa N AI+ is a Windows desktop built on NVIDIA's RTX Spark N1X, a 20-core Arm Grace CPU paired with a 6,144-core Blackwell RTX GPU, according to MSI's datasheet (read November 20, 2026). Until November 11, when MSI released it, that chip had shipped only in laptops. MSI first showed the machine at Computex in June 2026, StorageReview noted.

The memory is the change that matters for local models. MSI lists up to 128 GB of LPDDR5X unified memory, one pool that the CPU and GPU share, so a model is not limited to the size of a separate graphics card's memory. On StorageReview's unit, Windows gave the GPU 62.5 GB as dedicated memory by default, with another 47.9 GB shared. StorageReview notes that the memory cannot be upgraded later, so the capacity is chosen at purchase.

The cooler is the second change. MSI's datasheet says its vapour chamber and two turbo fans sustain a 140 W TDP for round-the-clock agent work, where the same chip in a laptop runs inside a thinner chassis and a battery budget. Power comes from a 240 W USB-C adapter, and the box measures 152 x 147 x 53 mm and weighs 1.32 kg, MSI lists.

The software stack is the third. The EdgeMesa runs Windows 11 on Arm with NVIDIA's CUDA stack, where its sibling, MSI's EdgeXpert, is a GB10 system that runs NVIDIA's Linux-based DGX OS and carries a 200 GbE ConnectX-7 port for clustering, StorageReview noted. The EdgeMesa has 10 Gb Ethernet and no clustering port.

The price point is the fourth. MSI's US store listed the 128 GB, 4 TB model at $7,999 when read on November 11, the figure our November 11 product file carried. On November 18, Notebookcheck reported that MSI had taken $500 off that model, to $7,499, and cut the 64 GB, 1 TB model from $4,699 to $4,399, after ASUS launched the GR1X at $4,499; Notebookcheck judged the cut permanent by the look of MSI's website. HotHardware's review, the same day, gave $4,399 as the starting price for a 64 GB configuration and $7,499 as the 128 GB version's MSRP. MSI's US store refused automated reads on November 20, so the $7,499 rests on Notebookcheck's and HotHardware's November 18 figures and on the MSRP NVIDIA gave StorageReview.

What it solves

The EdgeMesa lets a small team share one quiet desktop for models up to the 35B class without renting cloud GPUs, and StorageReview measured that directly. Its llama.cpp test on CUDA, at Q4_K_M quantization, sends one, four and eight simultaneous requests, which StorageReview describes as the load of a few people sharing one box. On Qwen3.8 27B, the EdgeMesa produced 13.0 tokens per second for one request, 31.1 for four and 39.8 for eight, with the first token arriving after 8.7 seconds.

The 128 GB of memory is what opened the larger tier. Running through Ollama on CUDA, the EdgeMesa loaded the 21 GB qwen3.6:35b model from cold in 8.32 seconds and passed StorageReview's coding-agent task with it, at a 64K context, on three of three attempts in 145 seconds at 86.8 tokens per second. The 128 GB HP OmniBook Ultra 16 laptop, on the same chip, passed one of three and took 1,783 seconds. The suite did not run that tier on 64 GB RTX Spark systems.

Image models gain the same way. Through stable-diffusion.cpp on Vulkan, the EdgeMesa ran Qwen-Image 20B at Q8 at 0.51 images per minute at 1024 x 1024, a version StorageReview did not run on its 64 GB machines.

The 120B class is still unproven on this box. MSI's product page, as StorageReview reports it, claims support for 120B-parameter models with up to 1 million tokens of context, a claim MSI's datasheet does not carry, but StorageReview said its Windows suite stops at the 35B agent model, so that claim went untested, and HotHardware published no text figures for model size.

Pros and Cons

Each item below rests on a measurement, a spec or a price from the sources named after the lists.

Concurrent serving is the clearest gain over laptops on the same chip. In StorageReview's llama.cpp test, the EdgeMesa's 39.8 tokens per second at eight requests on Qwen3.8 27B led the HP OmniBook Ultra 16 at 33.8 and the 64 GB Dell XPS 16 Creator Edition at 31.9, leads of 18 and 25 percent. With one request, all four RTX Spark systems landed within 6 percent of each other on every model. In MLPerf Client v2.0, which sends one prompt at a time, the EdgeMesa generated 56.3 tokens per second on Llama 3.1 8B, a four-way tie, StorageReview found.

Noise and power are where HotHardware's measurements speak. It wrote that the system did not register on its sound meter even under load for hours, that GPU temperatures peaked just under 80 C in 3DMark Steel Nomad Lite and Speed Way, and that Procyon AI's vision and text-generation tests ran between 50 C and 70 C with brief excursions into the low 80s. Its outlet meter showed peak system power in the 205+ W range in a gaming load. In HotHardware's AI tests (Procyon, computer vision and MLPerf Client), the EdgeMesa beat a Strix Halo small-form-factor system on every result; it gave those results in charts, and its text states only the ranking.

The storage is fast but closed. StorageReview measured the 4 TB Samsung Gen5 drive at 13,315 MB/s sequential read and 12,910 MB/s write, and a cold load of gpt-oss:20b in 7.00 seconds against 10.94 seconds on the GR1X's 1 TB Gen4 drive. But the drive is the short M.2 2242 size in the only slot, and HotHardware flagged low 4K random performance from it.

The CPU ran hot at its peaks. In StorageReview's 30-minute Cinebench 2026 loop, the EdgeMesa held 7,276 points with no throttling at an 88.4 C mean, but it peaked at 107.3 C and logged peaks above 100 C in three tests; the GR1X averaged 83.3 C in the same loop with an 88.7 C peak. StorageReview read a single 124.7 C sample in MLPerf Client as a sensor spike.

Ports and software are the other limits. HotHardware counted three usable USB-C ports, because the fourth takes power, and StorageReview found no USB Type-A port and no audio jack. HotHardware also raised Windows on Arm compatibility concerns and found that many MSI Center features, updates among them, were unavailable on this system.

Who it's for

A team of two to eight people sharing one local model in the 27B to 35B class is the best fit. StorageReview's eight-request result of 39.8 tokens per second on Qwen3.8 27B, and the 35B agent finishing three of three runs in 145 seconds, are the numbers behind that profile.

A developer who wants local agents and coding assistants inside Windows, rather than on a Linux box, is the second. MSI pitches the system at agents, retrieval and coding assistants, and StorageReview's RAG test with phi4-mini answered the median question in 0.23 seconds.

An office that wants a box running beside people for long sessions is the third. HotHardware wrote that its meter could not register the system's noise, even under load for hours.

It does not fit a buyer whose target is a 120B-class model at interactive speed today. No source has measured that on the EdgeMesa, and for a single user on a model of 8B or smaller, StorageReview found no speed difference from an RTX Spark laptop. A buyer who works only with models up to 27B can also spend less: StorageReview's 64 GB GR1X matched the 128 GB EdgeMesa's serving results within 2 percent, at $4,499.

How it compares

The ASUS ProArt GR1X is the direct rival on the same N1X chip. StorageReview measured its 64 GB unit at 13.1, 31.4 and 40.0 tokens per second at one, four and eight requests on Qwen3.8 27B, the same model, quantization and tool as the EdgeMesa's 13.0, 31.1 and 39.8. The GR1X ran 3 C to 8 C cooler on average in CPU tests and has two M.2 2280 slots, StorageReview found, while the EdgeMesa's 128 GB ran the 35B and Qwen-Image Q8 tiers. On paper, the two makers list different bandwidth for the same chip: MSI's US store gives 273 GB/s (read November 11), and ASUS's spec page gives 300 GB/s for the GR1X (read November 20). MSI's datasheet states no bandwidth figure, and HotHardware lists the EdgeMesa's memory as LPDDR5X-9600. StorageReview's CPU-side STREAM Triad test measured 122,423 MB/s on the EdgeMesa and 124,529 MB/s on the GR1X, which shows the two memory systems behave alike but does not test the GPU figure either maker lists. The GR1X starts at $4,499 with 64 GB and a 1 TB Gen4 drive, StorageReview reported.

On StorageReview's Qwen3.8 27B test the two machines land within 0.2 tok/s; memory and bandwidth are maker-listed figures.

Graphic: Voltgrid News

NVIDIA's DGX Spark, the Linux GB10 box, is the other 128 GB option from NVIDIA's platform. It lists 273 GB/s, and NVIDIA's marketplace showed the 128 GB, 4 TB model at $6,950 on October 31. No lab has published a DGX Spark result on the same model, quantization and tool as StorageReview's EdgeMesa runs. Our DGX Spark review covers its serving results and its clustering ports, which the EdgeMesa lacks.

AMD's Ryzen AI Halo is the price rival: StorageReview reported it at $3,999 with 128 GB and a 2 TB drive in July 2026. HotHardware called the EdgeMesa nearly double the price of that machine and found the EdgeMesa ahead of a Strix Halo small-form-factor system in every AI test, with figures shown only in charts. Our Ryzen AI Halo review has its measured results, taken on different models and tools from StorageReview's EdgeMesa runs.

Against the laptops on the same chip, covered in our RTX Spark laptop review, StorageReview put the EdgeMesa's $7,499 about $500 above the 128 GB XPS 16 Creator Edition at $6,999.99 and about $1,270 above the 128 GB OmniBook Ultra 16 at $6,229.99, prices from its reviews of those laptops, both of which shipped with 2 TB drives. HotHardware put the EdgeMesa a few hundred dollars above a Mac Studio with M5 Max, 128 GB and a 4 TB drive.

What to watch

Stock comes first. Notebookcheck reported on November 18 that the 128 GB model was out of stock at both Micro Center and MSI after the cut to $7,499, and MSI's US store had listed it as out of stock on November 11. Neither MSI nor the retailers had published a restock date by November 20.

The 64 GB model is the second. MSI's US price for it has moved from $4,699 to $4,399 in Notebookcheck's reporting, and no lab has published an LLM measurement of a 64 GB EdgeMesa; until one does, StorageReview's 64 GB GR1X is the nearest evidence on the same chip.

The 120B claim is the third. MSI's product page, as StorageReview reports it, advertises 120B-parameter models on the system, and StorageReview said it tests that class only in its Linux suite, so a Windows measurement at that size, from any lab, would settle the claim.

Software is the fourth. HotHardware expects MSI to bring MSI Center's update features to the Arm system at some point; MSI has not given a date. Notebookcheck's full GR1X review, which it promised on November 18, which it said would follow within days, should also add noise and power figures for the EdgeMesa's closest rival.

How it compares

Full table
MSI EdgeMesa N Review: 39.8 tok/s on a 27B Model, 35B Agent, Now $7,499: how the machine compares with other local AI hardware
ProductMemoryBandwidthAI computeMeasured speedPrice
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)MSIThis story128 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]
Show the other 34 machines
MSI EdgeMesa N Review: 39.8 tok/s on a 27B Model, 35B Agent, Now $7,499: 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]
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 GR1XASUS64 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]
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]
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

Changes

  1. Tom's Hardware published its review of the 128 GB EdgeMesa N AI+ on Nov 20. Under its AgentPerf-Local LLM test it measured about 140 W at the wall and 38.2 dBA at 18 inches (46 cm), and about 200 W and 40.7 dBA under gaming loads. On Qwen 3.8 27B with DFlash2 speculative decoding, the EdgeMesa performed practically the same as Microsoft's Surface Laptop Ultra, while AMD's Ryzen AI Halo, on Windows with the Vulkan backend, took about four times as long to finish. Tom's listed the price, limited ports and the lack of DGX Spark-style clustering as cons.

  2. First published.

Sources

Published
Updated
Sources
5 cited