can-i-run-this-llm

Can I run this LLM?

Can a NVIDIA L40S (48GB) run Llama 4 Scout (MoE)?

Llama · See model card · estimated for a 48.0 GB machine at 4K context.

Yes, but it's tight.

Best fit: Llama 4 Scout (MoE) at Q2_K · 39.6 GB · 73–109 tok/s on a NVIDIA L40S (48GB).

Every quantization on this machine

QuantSizeSpeedFit
Q2_K39.6 GB73–109 tok/sTight / slow
Q3_K_M51.8 GB60–89 tok/sWon't fit
Q4_K_M65.4 GB50–74 tok/sWon't fit
Q5_K_M76.6 GB44–65 tok/sWon't fit
Q6_K88.4 GB39–58 tok/sWon't fit
Q8_0114.5 GB31–46 tok/sWon't fit
FP16215.6 GB17–26 tok/sWon't fit

Ran Llama 4 Scout (MoE) on a NVIDIA L40S (48GB)? Add your real numbers →