can-i-run-this-llm

Can I run this LLM?

Can a NVIDIA RTX 3090 (24GB) run Phi-4 14B?

Phi · MIT · estimated for a 24.0 GB machine at 4K context.

Yes — it runs great.

Best fit: Phi-4 14B at Q8_0 · 15.6 GB · 38–57 tok/s on a NVIDIA RTX 3090 (24GB).

Every quantization on this machine

QuantSizeSpeedFit
Q2_K5.6 GB83–124 tok/sRuns great
Q3_K_M7.2 GB70–105 tok/sRuns great
Q4_K_M8.9 GB60–89 tok/sRuns great
Q5_K_M10.4 GB53–79 tok/sRuns great
Q6_K12.0 GB47–70 tok/sRuns great
Q8_015.6 GB38–57 tok/sRuns great
FP1629.3 GB22–32 tok/sWon't fit

Ran Phi-4 14B on a NVIDIA RTX 3090 (24GB)? Add your real numbers →