can-i-run-this-llm

Can I run this LLM?

Can a NVIDIA RTX 3090 (24GB) run Kimi Linear 48B-A3B (MoE)?

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

Yes — it runs great.

Best fit: Kimi Linear 48B-A3B (MoE) at Q2_K · 17.5 GB · 231–347 tok/s on a NVIDIA RTX 3090 (24GB).

Every quantization on this machine

QuantSizeSpeedFit
Q2_K17.5 GB231–347 tok/sRuns great
Q3_K_M22.7 GB209–314 tok/sWon't fit
Q4_K_M30.1 GB184–276 tok/sWon't fit
Q5_K_M35.1 GB170–255 tok/sWon't fit
Q6_K40.5 GB157–236 tok/sWon't fit
Q8_052.3 GB135–203 tok/sWon't fit
FP1698.3 GB87–131 tok/sWon't fit
Open the full calculator →Everything a NVIDIA RTX 3090 (24GB) runsKimi Linear 48B-A3B (MoE) on other machinesWhat to buy to run it →

Ran Kimi Linear 48B-A3B (MoE) on a NVIDIA RTX 3090 (24GB)? Add your real numbers →