can-i-run-this-llm

Can I run this LLM?

Can a AMD Instinct MI300X (192GB) run Hunyuan Hy3 (MoE)?

Hunyuan · Apache 2.0 · estimated for a 192.0 GB machine at 4K context.

Yes — it runs great.

Best fit: Hunyuan Hy3 (MoE) at Q3_K_M · 128.9 GB · 186–278 tok/s on a AMD Instinct MI300X (192GB).

Every quantization on this machine

QuantSizeSpeedFit
Q2_K97.1 GB208–311 tok/sRuns great
Q3_K_M128.9 GB186–278 tok/sRuns great
Q4_K_M181.1 GB158–237 tok/sWon't fit
Q5_K_M212.9 GB145–218 tok/sWon't fit
Q6_K246.5 GB133–200 tok/sWon't fit
Q8_0317.5 GB114–171 tok/sWon't fit

Ran Hunyuan Hy3 (MoE) on a AMD Instinct MI300X (192GB)? Add your real numbers →