can-i-run-this-llm

Can I run this LLM?

Can a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB) run GPT-OSS 20B (MoE)?

GPT-OSS · Apache 2.0 · estimated for a 144.0 GB machine at 4K context.

Yes — it runs great.

Best fit: GPT-OSS 20B (MoE) at FP16 · 13.8 GB · 291–436 tok/s on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB).

Every quantization on this machine

QuantSizeSpeedFit
Q2_K11.5 GB301–451 tok/sRuns great
Q3_K_M11.5 GB300–451 tok/sRuns great
Q4_K_M11.6 GB300–450 tok/sRuns great
Q5_K_M11.7 GB300–449 tok/sRuns great
Q6_K12.0 GB298–447 tok/sRuns great
Q8_012.1 GB298–447 tok/sRuns great
FP1613.8 GB291–436 tok/sRuns great

Ran GPT-OSS 20B (MoE) on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB)? Add your real numbers →