Yes — with 8.4 GB to spare
Llama 3.3 70B Instruct at Q4_K_M fits your CPU only · DDR5 dual-channel entirely on the GPU at 8K context, at an estimated 1.0 tokens per second. Past 34K the KV cache pushes it over — quantise the cache to q8_0, or step down a quantisation, to go longer.
Fully on GPU
8K context
Q4_K_M · 39.7 GB
Llama 3.3 Community
Released Dec 2024
Still the creative-writing favourite: consistent voice, takes direction. Needs 48 GB to sit comfortably on GPU at Q4.
The VRAM budget
weights 39.7 GB
Weights 39.7 GB
KV cache @ 8K 2.50 GB
Runtime overhead 0.6 GB
Free 8.4 GB of 51.2 GB
Quantisation ladder
| Quant | Weights | Total @ 8K | Max context | Tok/s | Quality | Fit |
|---|---|---|---|---|---|---|
| Q8_0 | 69.9 GB | 73.0 GB | — | ~0.5 | −0.1% ppl | 21.8 GB over |
| Q6_K | 53.9 GB | 57.0 GB | — | ~0.7 | −0.4% ppl | 5.8 GB over |
| Q5_K_M | 46.6 GB | 49.7 GB | 12K | 0.8 | −0.8% ppl | Fits |
| Q4_K_M | 39.7 GB | 42.8 GB | 34K | 1.0 | −1.9% ppl | Recommended |
| Q3_K_M | 32.1 GB | 35.2 GB | 59K | 1.2 | −5.4% ppl | Long context |
| Q2_K | 27.5 GB | 30.6 GB | 73K | 1.4 | −15% ppl | Long context |
Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it.
How to run it
$ llama-server \
-hf meta-llama/Llama-3.3-70B-Instruct:Q4_K_M \
-c 8192 -ngl 99
The engine underneath most of the others. Every knob is exposed. More on llama.cpp.
01Download is 39.7 GB. Keep it on an SSD — a first load off a spinning disk takes minutes.
02With no GPU, thread count matters more than clock. Start at one thread per physical core.
03There is room to go to 34K context on this card.