can-i-run-this-llm

Can I run this LLM?

Can a AMD Instinct MI300X (192GB) run Phi-4 14B?

Phi · MIT · estimated for a 192.0 GB machine at 4K context.

Yes — it runs great.

Best fit: Phi-4 14B at FP16 · 29.3 GB · 96–144 tok/s on a AMD Instinct MI300X (192GB).

Every quantization on this machine

QuantSizeSpeedFit
Q2_K5.6 GB229–343 tok/sRuns great
Q3_K_M7.2 GB209–314 tok/sRuns great
Q4_K_M8.9 GB192–288 tok/sRuns great
Q5_K_M10.4 GB179–268 tok/sRuns great
Q6_K12.0 GB166–249 tok/sRuns great
Q8_015.6 GB144–217 tok/sRuns great
FP1629.3 GB96–144 tok/sRuns great

Ran Phi-4 14B on a AMD Instinct MI300X (192GB)? Add your real numbers →