can-i-run-this-llm

Can I run this LLM?

Can a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB) run Qwen3.8 27B?

Qwen · Apache 2.0 · estimated for a 144.0 GB machine at 4K context.

Yes — it runs great.

Best fit: Qwen3.8 27B at FP16 · 54.7 GB · 56–84 tok/s on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB).

Every quantization on this machine

QuantSizeSpeedFit
Q2_K10.8 GB173–259 tok/sRuns great
Q3_K_M13.4 GB154–231 tok/sRuns great
Q4_K_M17.4 GB131–197 tok/sRuns great
Q5_K_M20.9 GB116–175 tok/sRuns great
Q6_K23.9 GB106–160 tok/sRuns great
Q8_029.1 GB92–138 tok/sRuns great
FP1654.7 GB56–84 tok/sRuns great

Ran Qwen3.8 27B on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB)? Add your real numbers →