can-i-run-this-llm

Can I run this LLM?

Can a NVIDIA RTX Pro 6000 Blackwell (96GB) run Llama 3.3 70B Instruct?

Llama · Llama 3.3 Community · estimated for a 96.0 GB machine at 4K context.

Yes — it runs great.

Best fit: Llama 3.3 70B Instruct at Q8_0 · 75.0 GB · 17–25 tok/s on a NVIDIA RTX Pro 6000 Blackwell (96GB).

Every quantization on this machine

QuantSizeSpeedFit
Q2_K26.4 GB42–63 tok/sRuns great
Q3_K_M34.3 GB34–51 tok/sRuns great
Q4_K_M42.5 GB28–42 tok/sRuns great
Q5_K_M50.0 GB24–36 tok/sRuns great
Q6_K57.9 GB21–32 tok/sRuns great
Q8_075.0 GB17–25 tok/sRuns great
FP16141.1 GB9–14 tok/sWon't fit

Ran Llama 3.3 70B Instruct on a NVIDIA RTX Pro 6000 Blackwell (96GB)? Add your real numbers →