Can I Run Magistral Small on Framework Desktop (Ryzen AI Max+ 395, 128 GB)?

Written by Jakub Rusinowski · Last updated September 6, 2026

Yes — comfortably

Yes, comfortably — Magistral Small 24B at Q8_0 needs about 28 GB of the 128 GB usable on Framework Desktop (Ryzen AI Max+ 395, 128 GB), leaving ~100 GB spare and running at ~7.2 tok/s (estimated), with room for about 131,072 tokens of context.

Confidence: medium · Recommended quantization: Q8_0 · Estimated speed: ~7.2 tok/s

Buy This HardwareApple Mac mini M4 (16GB) — 32 GB VRAM · 22 W board power Deploy in the Cloud NowNVIDIA A40 on RunPod — from $0.44/hr · rate checked 2026-08

or compare on Vast.ai

As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.

Framework Desktop (Ryzen AI Max+ 395, 128 GB) — what it gives a model

Usable memory for models	128 GB
Memory bandwidth	256 GB/s
Form factor	Mini PC
Operating system	Windows or Linux
Memory upgradeable	No — soldered

Magistral Small on Framework Desktop (Ryzen AI Max+ 395, 128 GB): memory by quantization

Quant	Memory needed	Fits 128 GB?	Max context	Est. speed	Download
F16	50.5 GB	✓ Yes	128K	~3.9 tok/s	48 GB
Q8_0	28 GB	✓ Yes	128K	~7.2 tok/s	25.5 GB
Q6_K	22.2 GB	✓ Yes	128K	~9.2 tok/s	19.7 GB
Q5_K_M	19.5 GB	✓ Yes	128K	~10.5 tok/s	17 GB
Q4_K_M	17 GB	✓ Yes	128K	~12.2 tok/s	14.5 GB
Q3_K_M	12.7 GB	✓ Yes	128K	~16.7 tok/s	10.2 GB
Q2_K	10.4 GB	✓ Yes	128K	~21 tok/s	7.9 GB

What to watch out for

This model does not publish its full attention configuration, so the KV-cache share of these figures is inferred from its parameter count rather than computed exactly.
Memory on this machine is not upgradeable, so the configuration you buy is the ceiling for every model you will ever run on it.

Framework Desktop 128 GB limitations

Memory is soldered LPDDR5X — unusually for Framework, this is the one component that cannot be upgraded.
ROCm/Vulkan support for Strix Halo is younger than CUDA; check your runtime supports it before buying.

Recommended setup

Ollama or llama.cpp (CUDA/ROCm) — vLLM if you need concurrent serving

How these numbers are calculated

Memory need = quantized weights + KV cache at 8,192 tokens (f16) + 0.8 GB runtime overhead.
128 GB unified memory at 256 GB/s, shared between CPU and GPU.
Throughput is a memory-bandwidth roofline estimate, not a measurement. It is labelled "estimated" everywhere it appears.
KV cache is inferred from the parameter count because this model does not publish its full attention configuration.

FAQ