Can I run this LLM?
Can a NVIDIA GB300 Grace Blackwell (288GB HBM + 480GB) run Kimi Linear 48B-A3B (MoE)?
Kimi · MIT · estimated for a 288.0 GB machine at 4K context.
Yes — it runs great.
Best fit: Kimi Linear 48B-A3B (MoE) at FP16 · 98.2 GB · 267–400 tok/s on a NVIDIA GB300 Grace Blackwell (288GB HBM + 480GB).
Every quantization on this machine
| Quant | Size | Speed | Fit |
|---|---|---|---|
| Q2_K | 16.0 GB | 345–517 tok/s | Runs great |
| Q3_K_M | 21.2 GB | 338–508 tok/s | Runs great |
| Q4_K_M | 29.8 GB | 329–493 tok/s | Runs great |
| Q5_K_M | 35.0 GB | 323–484 tok/s | Runs great |
| Q6_K | 40.5 GB | 317–475 tok/s | Runs great |
| Q8_0 | 52.2 GB | 305–458 tok/s | Runs great |
| FP16 | 98.2 GB | 267–400 tok/s | Runs great |
Open the full calculator →Everything a NVIDIA GB300 Grace Blackwell (288GB HBM + 480GB) runsKimi Linear 48B-A3B (MoE) on other machinesWhat to buy to run it →
Ran Kimi Linear 48B-A3B (MoE) on a NVIDIA GB300 Grace Blackwell (288GB HBM + 480GB)? Add your real numbers →