can-i-run-this-llm

Can I run this LLM?

Can a AMD Instinct MI300X (192GB) run Llama 4 Scout (MoE)?

Llama · See model card · estimated for a 192.0 GB machine at 4K context.

Yes — it runs great.

Best fit: Llama 4 Scout (MoE) at Q8_0 · 114.5 GB · 133–199 tok/s on a AMD Instinct MI300X (192GB).

Every quantization on this machine

QuantSizeSpeedFit
Q2_K39.6 GB221–331 tok/sRuns great
Q3_K_M51.8 GB199–299 tok/sRuns great
Q4_K_M65.4 GB180–270 tok/sRuns great
Q5_K_M76.6 GB166–249 tok/sRuns great
Q6_K88.4 GB154–231 tok/sRuns great
Q8_0114.5 GB133–199 tok/sRuns great
FP16215.6 GB86–129 tok/sWon't fit

Ran Llama 4 Scout (MoE) on a AMD Instinct MI300X (192GB)? Add your real numbers →