Can I Run Granite 3.0 on Framework Desktop (Ryzen AI Max+ 395, 128 GB)?

Superseded model. Granite 3.0 has been superseded by IBM Granite 4.0. This page is kept for reference; the newer family is a better starting point. View IBM Granite 4.0 →

Written by Jakub Rusinowski · Last updated October 21, 2024

Yes — comfortably

Yes, comfortably — Granite 3.0 8B Instruct at Q8_0 needs about 10.6 GB of the 128 GB usable on Framework Desktop (Ryzen AI Max+ 395, 128 GB), leaving ~117.4 GB spare and running at ~20 tok/s (estimated), with room for about 131,072 tokens of context.

Confidence: high · Recommended quantization: Q8_0 · Estimated speed: ~20 tok/s

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

Usable memory for models128 GB
Memory bandwidth256 GB/s
Form factorMini PC
Operating systemWindows or Linux
Memory upgradeableNo — soldered

Granite 3.0 on Framework Desktop (Ryzen AI Max+ 395, 128 GB): memory by quantization

QuantMemory neededFits 128 GB?Max contextEst. speedDownload
F1618.1 GB✓ Yes128K~11.2 tok/s16 GB
Q8_010.6 GB✓ Yes128K~20 tok/s8.5 GB
Q6_K8.7 GB✓ Yes128K~25.1 tok/s6.6 GB
Q5_K_M7.8 GB✓ Yes128K~28.4 tok/s5.7 GB
Q4_K_M7 GB✓ Yes128K~32.4 tok/s4.8 GB
Q3_K_M5.6 GB✓ Yes128K~42.6 tok/s3.4 GB
Q2_K4.8 GB✓ Yes128K~51.4 tok/s2.6 GB

What to watch out for

Framework Desktop 128 GB limitations

Recommended setup

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

How these numbers are calculated

FAQ

Can I run Granite 3.0 on Framework Desktop (Ryzen AI Max+ 395, 128 GB)?

Yes, comfortably — Granite 3.0 8B Instruct at Q8_0 needs about 10.6 GB of the 128 GB usable on Framework Desktop (Ryzen AI Max+ 395, 128 GB), leaving ~117.4 GB spare and running at ~20 tok/s (estimated), with room for about 131,072 tokens of context.

Which quantization of Granite 3.0 should I use on Framework Desktop (Ryzen AI Max+ 395, 128 GB)?

Q8_0 — it needs about 10.6 GB of the 128 GB available, downloads as roughly 8.5 GB, and runs at an estimated 20 tokens/sec with up to 128K of context.

What limits Granite 3.0 on Framework Desktop (Ryzen AI Max+ 395, 128 GB)?

Nothing binding — the model fits with headroom and generates at a usable speed on this hardware.

Which runtime should I use?

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

Other Computers

Other Models on Framework Desktop (Ryzen AI Max+ 395, 128 GB)

Granite 3.0 on GPUs

What This Model Is Good At

Model & Tools

← Can I Run It? | Granite 3.0 model page | Check your hardware