Can I run Llama 3.1 70B on an RTX 3060 12GB?
No: Llama 3.1 70B needs about 55.2 GB at Q4_K_M with 32K tokens of context, 43.2 GB more than the RTX 3060 12GB holds, so it takes 8 of them (96 GB) in one tensor-parallel group. The smallest setup here that holds Llama 3.1 70B at Q4_K_M with 32K is 2× RTX 5090 (64 GB).
No Q4_K_M with 32K tokens of context on 8 cards
- Q4_K_M, 32K
- 55.2 GB
- RTX 3060
- 12 GB, 360 GB/s
- Short by
- 43.2 GB
- Tokens/s
- 27–38 tokens/s
At Q4_K_M with 32K tokens of context on 8 cards it writes about 27–38 tokens/s for one request on 8 RTX 3060 12GB cards.
Llama 3.1 70B on the RTX 3060 as the context fills
One request, FP16 KV cache, 0.5 GB plus 10% overhead; a minus sign is memory missing, and tight is less than 0.5 GB free.
| Context | Q4_K_M | Free | Tokens/s | Q8_0 | Free | Tokens/s |
|---|---|---|---|---|---|---|
| 4K | 45.6 GB | −33.6 GB | — | 78.7 GB | −66.7 GB | — |
| 8K | 47.0 GB | −35.0 GB | — | 80.0 GB | −68.0 GB | — |
| 16K | 49.7 GB | −37.7 GB | — | 82.8 GB | −70.8 GB | — |
| 32K | 55.2 GB | −43.2 GB | — | 88.3 GB | −76.3 GB | — |
| 64K | 66.2 GB | −54.2 GB | — | 99.3 GB | −87.3 GB | — |
| 128K (full) | 88.2 GB | −76.2 GB | — | 121 GB | −109 GB | — |
Llama 3.1 70B on more than one RTX 3060
| Cards | Best at 32K | Tokens/s | Q4_K_M longest | Q4_K_M tokens/s | Rent per hour |
|---|---|---|---|---|---|
| 2× (24 GB) | Nothing fits | — | Does not fit | — | $0.16 |
| 4× (48 GB) | Q2_K | 18–26 | 4K | — | $0.32 |
Tensor parallel with the combined bandwidth, as in vLLM; llama.cpp splits layers by default and runs at about one card's speed.
Other options
- The next smaller setting, Q3_K_M, takes 46.8 GB at 32K, 34.8 GB over the RTX 3060; it does not fit with 0.5 GB to spare even with 1K tokens.
- Renting eight RTX 3060 12GB GPUs costs about $0.64 an hour (median on getdeploying.com, 2026-09-29): $4.6–6.6 per million tokens at 27–38 tokens/s.
- Smallest setup for Q4_K_M at 32K: 2× RTX 5090 (64 GB).
Run Llama 3.1 70B on the RTX 3060 with llama-server
llama-server -hf bartowski/Meta-Llama-3.1-70B-Instruct-GGUF:Q4_K_M -c 131072Meta-Llama-3.1-70B-Instruct-Q4_K_M.gguf, 42.5 GB, from bartowski/
Questions
Can I run Llama 3.1 70B on an RTX 3060 12GB?
No: Llama 3.1 70B needs about 55.2 GB at Q4_K_M with 32K tokens of context, 43.2 GB more than the RTX 3060 12GB holds, so it takes 8 of them (96 GB) in one tensor-parallel group. The smallest setup here that holds Llama 3.1 70B at Q4_K_M with 32K is 2× RTX 5090 (64 GB).
How fast is Llama 3.1 70B on an RTX 3060 12GB?
At Q4_K_M with 32K tokens of context on 8 cards it writes about 27–38 tokens/s for one request on 8 RTX 3060 12GB cards.
What does a second RTX 3060 12GB change for Llama 3.1 70B?
Even two RTX 3060 12GB cards (24 GB) do not hold Llama 3.1 70B at 32K.
Try other settings in the VRAM calculator or the speed calculator. See also Llama 3.1 70B VRAM requirements, what LLMs an RTX 3060 12GB can run and every pair, or detect your own GPU. Model data checked .