This page answers RedHatAI/Meta-Llama-3.1-70B-Instruct-quantized.w4a16 q4 quantization queries with explicit calculations from our model requirement dataset and compatibility speed table.
Exact Q4 requirement from model requirement data.
Throughput data below uses available compatibility measurements/estimates and is sorted by tokens per second for this model.
Need general guidance? Review full methodology.
| GPU | VRAM | Quantization | Speed | Compatibility |
|---|---|---|---|---|
| AMD Instinct MI300X | 192GB | Q4 | 276 tok/s | View full compatibility |
| NVIDIA H200 SXM 141GB | 141GB | Q4 | 234 tok/s | View full compatibility |
| AMD Instinct MI250X | 128GB | Q4 | 175 tok/s | View full compatibility |
| NVIDIA H100 SXM5 80GB | 80GB | Q4 | 174 tok/s | View full compatibility |
| RTX 5090 | 32GB | Q4 | 107 tok/s | View full compatibility |
| NVIDIA H100 PCIe 80GB | 80GB | Q4 | 106 tok/s | View full compatibility |
| NVIDIA A100 80GB SXM4 | 80GB | Q4 | 105 tok/s | View full compatibility |
| NVIDIA A100 40GB PCIe | 40GB | Q4 | 87 tok/s | View full compatibility |
| AMD Instinct MI210 | 64GB | Q4 | 83 tok/s | View full compatibility |
| NVIDIA L40 | 48GB | Q4 | 64 tok/s | View full compatibility |
| NVIDIA RTX 6000 Ada | 48GB | Q4 | 62 tok/s | View full compatibility |
| RTX 4090 | 24GB | Q4 | 62 tok/s | View full compatibility |