can-i-run-this-llm

Can I run this LLM?

Can a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB) run DeepSeek-R1-Distill 8B?

DeepSeek · Llama 3.1 Community · estimated for a 144.0 GB machine at 4K context.

Yes — it runs great.

Best fit: DeepSeek-R1-Distill 8B at FP16 · 16.1 GB · 137–205 tok/s on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB).

Every quantization on this machine

QuantSizeSpeedFit
Q2_K3.2 GB267–400 tok/sRuns great
Q3_K_M4.0 GB251–376 tok/sRuns great
Q4_K_M4.9 GB236–354 tok/sRuns great
Q5_K_M5.7 GB224–337 tok/sRuns great
Q6_K6.6 GB213–319 tok/sRuns great
Q8_08.5 GB191–287 tok/sRuns great
FP1616.1 GB137–205 tok/sRuns great

Ran DeepSeek-R1-Distill 8B on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB)? Add your real numbers →