can-i-run-this-llm

Can I run this LLM?

Can a Mac Studio M3 Ultra (96GB) run Ornith 1.0 35B (MoE)?

Ornith · MIT · estimated for a 96.0 GB machine at 4K context.

Yes — it runs great.

Best fit: Ornith 1.0 35B (MoE) at Q8_0 · 36.9 GB · 82–123 tok/s on a Mac Studio M3 Ultra (96GB).

Every quantization on this machine

QuantSizeSpeedFit
Q3_K_M16.7 GB111–166 tok/sRuns great
Q4_K_M22.1 GB101–152 tok/sRuns great
Q5_K_M26.5 GB95–142 tok/sRuns great
Q6_K29.3 GB91–136 tok/sRuns great
Q8_036.9 GB82–123 tok/sRuns great
FP1669.4 GB58–87 tok/sWon't fit

Ran Ornith 1.0 35B (MoE) on a Mac Studio M3 Ultra (96GB)? Add your real numbers →