Llama 3.1 8B hardware requirements
8.03B parameters. Needs about 6.5 GB at Q4_K_M with 8K context. Model card.
Memory needed at each quantization
| Quant | Quality | 8K context | 32K context |
|---|---|---|---|
| FP16 | Full quality | 17.7 GB | 21 GB |
| Q8_0 | Near-lossless | 10.2 GB | 13.4 GB |
| Q6_K | Excellent | 8.3 GB | 11.5 GB |
| Q5_K_M | Very good | 7.4 GB | 10.6 GB |
| Q4_K_M | Good, most popular | 6.5 GB | 9.8 GB |
| Q3_K_M | Noticeably degraded | 5.6 GB | 8.8 GB |
| Q2_K | Heavily degraded | 4.7 GB | 7.9 GB |
Run it with Ollama: ollama run llama3.1:8b (downloads Q4_K_M by default).
Run it in the cloud
Don't want to buy hardware? Run Llama 3.1 8B on a rented GPU today.
Rent a cloud GPU by the hour. No upfront cost, cancel any time.
Every GPU and Mac checked
| GPU | Runs? | Best fit |
|---|---|---|
| NVIDIA RTX 3060 12GB | Yes | Q8_0 · ~10.2 GB · ~25 tok/s |
| NVIDIA RTX 3070 | Yes | Q5_K_M · ~7.4 GB · ~47 tok/s |
| NVIDIA RTX 3080 10GB | Yes | Q6_K · ~8.3 GB · ~69 tok/s |
| NVIDIA RTX 3090 | Yes | FP16 · ~17.7 GB · ~35 tok/s |
| NVIDIA RTX 4060 | Yes | Q5_K_M · ~7.4 GB · ~29 tok/s |
| NVIDIA RTX 4060 Ti 16GB | Yes | Q8_0 · ~10.2 GB · ~20 tok/s |
| NVIDIA RTX 5060 Ti 16GB | Yes | Q8_0 · ~10.2 GB · ~32 tok/s |
| NVIDIA RTX 4070 | Yes | Q8_0 · ~10.2 GB · ~35 tok/s |
| NVIDIA RTX 4070 Ti Super | Yes | Q8_0 · ~10.2 GB · ~47 tok/s |
| NVIDIA RTX 4080 | Yes | Q8_0 · ~10.2 GB · ~50 tok/s |
| NVIDIA RTX 4090 | Yes | FP16 · ~17.7 GB · ~38 tok/s |
| NVIDIA RTX 5070 | Yes | Q8_0 · ~10.2 GB · ~47 tok/s |
| NVIDIA RTX 5070 Ti | Yes | Q8_0 · ~10.2 GB · ~63 tok/s |
| NVIDIA RTX 5080 | Yes | Q8_0 · ~10.2 GB · ~68 tok/s |
| NVIDIA RTX 5090 | Yes | FP16 · ~17.7 GB · ~67 tok/s |
| AMD Radeon RX 7800 XT | Yes | Q8_0 · ~10.2 GB · ~44 tok/s |
| AMD Radeon RX 7900 XTX | Yes | FP16 · ~17.7 GB · ~36 tok/s |
| Apple M4 (16GB) | Yes | Q8_0 · ~10.2 GB · ~8 tok/s |
| Apple M1 (16GB) | Yes | Q8_0 · ~10.2 GB · ~5 tok/s |
| Apple M2 Pro (32GB) | Yes | FP16 · ~17.7 GB · ~7 tok/s |
| Apple M3 Max (64GB) | Yes | FP16 · ~17.7 GB · ~15 tok/s |
| Apple M4 Pro (48GB) | Yes | FP16 · ~17.7 GB · ~10 tok/s |
| Apple M4 Max (128GB) | Yes | FP16 · ~17.7 GB · ~20 tok/s |
| Apple M2 Ultra (192GB) | Yes | FP16 · ~17.7 GB · ~30 tok/s |