can-i-run-this-llm

Can I run this LLM?

Can a NVIDIA L40S (48GB) run Ornith 1.5 35B-A3B (MoE)?

Ornith · MIT · estimated for a 48.0 GB machine at 4K context.

Yes — it runs great.

Best fit: Ornith 1.5 35B-A3B (MoE) at Q8_0 · 37.8 GB · 127–191 tok/s on a NVIDIA L40S (48GB).

Every quantization on this machine

QuantSizeSpeedFit
Q2_K13.1 GB217–326 tok/sRuns great
Q3_K_M16.7 GB197–295 tok/sRuns great
Q4_K_M21.9 GB174–260 tok/sRuns great
Q5_K_M25.5 GB160–240 tok/sRuns great
Q6_K30.5 GB145–217 tok/sRuns great
Q8_037.8 GB127–191 tok/sRuns great
FP1671.1 GB81–122 tok/sWon't fit

Ran Ornith 1.5 35B-A3B (MoE) on a NVIDIA L40S (48GB)? Add your real numbers →