Home›Hardware›PC with Ryzen AI Max+ 395 and 128 GB unified

Hardware sheet

PC with Ryzen AI Max+ 395 and 128 GB unified

AMD chip with unified memory128 GiB256 GB/s14.8 TFLOPS

Framework Desktop, one of the PCs with this chip
Framework Desktop, one of the PCs with this chip Photo: Framework
Minisforum MS-S1 MAX, another PC with this chip
Minisforum MS-S1 MAX, another PC with this chip Photo: Minisforum

Watch 2 videos

Once the system takes its share, this machine leaves 96.00 GiB for the model: room for 24 of the 28 models we measure at the reference compression, and the heaviest one that fits is qwen3.8-flash-next. And how fast does it answer? That’s not down to the chip, but to its 256 GB/s of memory bandwidth.

At a glanceMachine data · 2026-10-30
Left for the model96.00 GiB
Memory bandwidth256 GB/s
Models that fit24 / 28
Heaviest model it takesqwen3.8-flash-next
96.00 GiB

Left for the model

Estimated

256 GB/s

Memory bandwidth

Estimated

24 / 28

Models that fit

Measured by Local AI Scope

Videos

Videos: PC with Ryzen AI Max+ 395 and 128 GB unified 2

The official videos we found most useful. Press play and it opens right here.

The videos live on YouTube, and nothing loads from there until you press play (no-cookie mode). The thumbnail is served by us.

Memory

The memory you can actually use

Unified memory is shared with the operating system and everything you have open. The reserve here is deliberately conservative; you can change it in the tool.

How the memory on the box turns into memory you can use
Memory GiB
Memory on the box 128
Reserved for the system −3.00
Beyond the limit the manufacturer sets for the GPU −29.00
Left for the model 96.00
What the PC with Ryzen AI Max+ 395 and 128 GB unified really has free for a modelHorizontal bar of the 128 GiB of the PC with Ryzen AI Max+ 395 and 128 GB unified, split into 32.0 GiB outside what the GPU can use, 87.2 GiB for the qwen3.8-flash-next model, 0.9 GiB for the context and 7.8 GiB free.0128 GiBOutside GPU: 32.0 GiBOutside GPU32.0Model: 87.2 GiBModel87.2Context (8,192): 0.9 GiBContext (8,192) 0.9Headroom: 7.8 GiBHeadroom 7.8
Of the 128 GiB the machine has, 32.0 GiB never reach the model. The bar shows qwen3.8-flash-next at IQ4_XS, the largest of the 28 models in the dataset that still fits at a 8,192-token context, and the KV cache that context needs — computed layer by layer, not with the usual rule of thumb.

128 GB on the box, 96 for the model. AMD states that on this generation up to 96 GB of the 128 GB are reserved for the graphics processor, and the graphics processor is what runs the model. So every figure on this page is computed with 96, not with the 125 that the box minus our system reserve would give. On Linux that limit is a driver setting that AMD documents as adjustable, but it publishes no maximum figure for it: we stick to the figure it does publish.

Speed

Memory bandwidth

Two different limits, and they are easy to mix up. While the model is writing its answer, bandwidth rules: every token means reading the weights it uses again — all of them in a dense model; in a mixture of experts, the shared layers plus only the active experts — so tokens per second is roughly bandwidth divided by that amount. While the model is reading your document, compute rules: the whole input is processed at once. That is why a laptop with no GPU can take minutes before the first word appears and then type at a tolerable pace.

256 GB/s — Derived, not published by the manufacturer · the manufacturer publishes the bus width but neither the memory speed nor the bandwidth. This figure is the usual arithmetic derivation () and we could not confirm it at source (Source).

Compute: 14.8 TFLOPS (estimate for this class of machine) — Estimated · the manufacturer does not publish this figure at all; ours is a conservative estimate, not a specification. The figures in this line do not all come from the same measure across machines, because each manufacturer publishes a different one. Compare them between machines only with that in mind.

Fit

Which models fit here, and which do not

Fitting means three things at once fit in the memory left over: the model file, its context memory, and the runtime’s working headroom. Context memory is computed from each model’s declared attention pattern layer by layer, not from the generic formula — which is why several models fit here that the generic formula would rule out.

Every measured model against this machine: size, whether it fits at each context, and the largest context it holds
Model Compression shown Model size 8k 32k 128k Largest context that fits
qwen3-1.7b Q4_K_M 1.03 GiB yes yes — 32k
qwen3-4b-2507 Q4_K_M 2.33 GiB yes yes yes 256k
gemma4-e2b Q4_K_M 2.89 GiB yes yes yes 128k
gemma4-e4b Q4_K_M 4.63 GiB yes yes yes 128k
qwen3-8b Q4_K_M 4.68 GiB yes yes — 32k
granite-4.2-8b Q4_K_M 4.98 GiB yes yes yes 128k
qwen3.5-9b Q4_K_M 5.29 GiB yes yes yes 256k
gemma4-12b Q4_K_M 6.63 GiB yes yes yes 256k
phi-4 Q4_K_M 8.43 GiB yes — — 16k
gpt-oss-20b Q4_K_M 10.83 GiB yes yes yes 128k
mistral-small-24b Q4_K_M 13.35 GiB yes yes yes 128k
qwen3.8-27b Q4_K_M 15.33 GiB yes yes yes 256k
muse-glimmer-30b Q4_K_M 15.61 GiB yes yes yes 128k
qwen3.6-27b Q4_K_M 15.66 GiB yes yes yes 256k
gemma4-26b-a4b Q4_K_M 15.78 GiB yes yes yes 256k
gemma4-31b Q4_K_M 17.07 GiB yes yes yes 256k
qwen3-coder-30b Q4_K_M 17.28 GiB yes yes yes 256k
qwen3.6-35b-a3b Q4_K_M 20.61 GiB yes yes yes 256k
flint-40b Q4_K_M 23.61 GiB yes yes yes 256k
nemotron-3.5-lightning-30b Q4_K_M 23.73 GiB yes yes yes 256k
gpt-oss-120b Q4_K_M 58.46 GiB yes yes yes 128k
llama4-scout-17b Q4_K_M 60.87 GiB yes yes yes 256k
nemotron-3.5-super-120b Q4_K_M 65.29 GiB yes yes yes 1024k
qwen3.8-flash-next IQ4_XS 87.25 GiB yes yes yes 128k
deepseek-v4-flash IQ4_XS 127.28 GiB no no no —
glm-5.3-flash Q4_K_M 187.04 GiB no no no —
glm-5.2 Q4_K_M 433.83 GiB no no no —
mistral-large-4 Q4_K_M 591.21 GiB no no no —

Sizes are the real published files at the reference compression, one step per row: Q4_K_M where the author publishes it, the nearest neighbour where they do not. Every row states which one it is showing. Q4_K_M is our quality floor: below it the loss is audible in the answers. A model that would only fit here at a harsher compression is listed as not fitting, on purpose. A dash in a context column means the model itself does not offer that context, so there is nothing to fit.

Models that fit — 24 of the 28 models measured

Models that do not fit — 4 of the 28 models measured

The machines on either side

Price

Price and availability

What a machine costs decides as much as what fits inside it, so the price belongs on this page. What follows is only what the manufacturer publishes, with the market it applies to and the day it was read.

From $3,449 (United States), as published on 2026-10-09 (Source). Vendor declared

That is the Framework Desktop with the Max+ 395 and 128 GB, before tax, and on 30 October 2026 every configuration was out of stock (in Framework’s German store: €3,889, VAT included). The one we found in stock that day with the same chip and 128 GB is GMKtec’s EVO-X2 with a 1 TB drive: $3,499.99 in its own store, €3,349.99 in its European one. Every figure on this page is computed for the chip and its 128 GB, so it holds for both.

What we earn on this page

Nothing, today. There is no affiliate link on this page and no commercial agreement behind any figure on it. If a buying link ever appears here, it will be marked as such and the commission declared in this same block. Two rules do not change on that day: the ranking on this site is computed before commercial availability is applied, and «best» never means «pays most».

Our affiliate policy