can-i-run-this-llm

Can I run this LLM?

Can a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB) run Hunyuan Hy3 (MoE)?

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

Yes — it runs great.

Best fit: Hunyuan Hy3 (MoE) at Q8_0 · 317.5 GB · 19–28 tok/s on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB).

Every quantization on this machine

QuantSizeSpeedFit
Q2_K97.1 GB201–301 tok/sRuns great
Q3_K_M128.9 GB169–253 tok/sRuns great
Q4_K_M181.1 GB53–79 tok/sRuns great
Q5_K_M212.9 GB37–56 tok/sRuns great
Q6_K246.5 GB28–43 tok/sRuns great
Q8_0317.5 GB19–28 tok/sRuns great

Ran Hunyuan Hy3 (MoE) on a NVIDIA GH200 Grace Hopper (144GB HBM + 480GB)? Add your real numbers →