can-i-run-this-llm

Can I run this LLM?

Can a MacBook Pro M3 Pro (36GB) run Kimi Linear 48B-A3B (MoE)?

Kimi · MIT · estimated for a 36.0 GB machine at 4K context.

Yes — it runs great.

Best fit: Kimi Linear 48B-A3B (MoE) at Q3_K_M · 21.2 GB · 52–78 tok/s on a MacBook Pro M3 Pro (36GB).

Every quantization on this machine

QuantSizeSpeedFit
Q2_K16.0 GB62–93 tok/sRuns great
Q3_K_M21.2 GB52–78 tok/sRuns great
Q4_K_M29.8 GB41–61 tok/sWon't fit
Q5_K_M35.0 GB36–54 tok/sWon't fit
Q6_K40.5 GB32–48 tok/sWon't fit
Q8_052.2 GB26–39 tok/sWon't fit
FP1698.2 GB15–23 tok/sWon't fit

Ran Kimi Linear 48B-A3B (MoE) on a MacBook Pro M3 Pro (36GB)? Add your real numbers →