can-i-run-this-llm

Can I run this LLM?

Can a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB) run Ornith 1.5 35B-A3B (MoE)?

Ornith · MIT · estimated for a 144.0 GB machine at 4K context.

Yes — it runs great.

Best fit: Ornith 1.5 35B-A3B (MoE) at FP16 · 71.1 GB · 227–340 tok/s on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB).

Every quantization on this machine

QuantSizeSpeedFit
Q2_K13.1 GB327–491 tok/sRuns great
Q3_K_M16.7 GB318–478 tok/sRuns great
Q4_K_M21.9 GB307–460 tok/sRuns great
Q5_K_M25.5 GB299–448 tok/sRuns great
Q6_K30.5 GB289–433 tok/sRuns great
Q8_037.8 GB275–413 tok/sRuns great
FP1671.1 GB227–340 tok/sRuns great

Ran Ornith 1.5 35B-A3B (MoE) on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB)? Add your real numbers →