can-i-run-this-llm

Can I run this LLM?

Can a AMD Ryzen AI Max+ 395 “Strix Halo” (128GB) run Qwen3.8 27B?

Qwen · Apache 2.0 · estimated for a 128.0 GB machine at 4K context.

Yes — it runs great.

Best fit: Qwen3.8 27B at Q2_K · 10.8 GB · 14–21 tok/s on a AMD Ryzen AI Max+ 395 “Strix Halo” (128GB).

Every quantization on this machine

QuantSizeSpeedFit
Q2_K10.8 GB14–21 tok/sRuns great
Q3_K_M13.4 GB12–18 tok/sTight / slow
Q4_K_M17.4 GB9–14 tok/sTight / slow
Q5_K_M20.9 GB8–12 tok/sTight / slow
Q6_K23.9 GB7–10 tok/sTight / slow
Q8_029.1 GB6–9 tok/sTight / slow
FP1654.7 GB3–5 tok/sTight / slow

Ran Qwen3.8 27B on a AMD Ryzen AI Max+ 395 “Strix Halo” (128GB)? Add your real numbers →