GPU Requirements
VRAM / RAM / disk estimates by quantization and usage scenario
| Quantization | Minimum | Recommended | Production |
|---|---|---|---|
| FP1616bit | 174.6 GB RAM 261.9 · Disk 248.2 | 198.1 GB RAM 297.1 · Disk 274.2 | 302.7 GB RAM 454.0 · Disk 313.2 |
| FP88bit | 89.8 GB RAM 134.7 · Disk 143.6 | 109.4 GB RAM 164.1 · Disk 169.6 | 210.1 GB RAM 315.2 · Disk 208.6 |
| INT88bit | 89.8 GB RAM 134.7 · Disk 143.6 | 109.4 GB RAM 164.1 · Disk 169.6 | 210.1 GB RAM 315.2 · Disk 208.6 |
| AWQ4bit | 49.0 GB RAM 73.4 · Disk 93.3 | 66.7 GB RAM 100.1 · Disk 119.3 | 165.6 GB RAM 248.4 · Disk 158.3 |
| GPTQ4bit | 49.5 GB RAM 74.2 · Disk 93.9 | 67.3 GB RAM 100.9 · Disk 119.9 | 166.2 GB RAM 249.3 · Disk 158.9 |
| GGUF4bit | 50.5 GB RAM 75.8 · Disk 95.2 | 68.4 GB RAM 102.6 · Disk 121.2 | 167.3 GB RAM 251.0 · Disk 160.2 |
GPU Recommendations
Best ValueBEST
1× H200 SXM
141 GB VRAM
$2,555/mo
≈ 138 tok/s
Effective production throughput ≈ 96.6 tok/s
score 0.5972
Cheapest
8× Tesla V100
128 GB VRAM
$169/mo
≈ 0 tok/s (N/A — benchmark unavailable)
Effective production throughput ≈ 0 tok/s
score 0.4954
Performance
8× H200 SXM
1128 GB VRAM
$20,440/mo
≈ 441.6 tok/s
Effective production throughput ≈ 309.1 tok/s
score 0.2197
Min Complexity
1× A100 80GB
80 GB VRAM
$876/mo
≈ 58.6 tok/s
Effective production throughput ≈ 41 tok/s
score 0.664
💬 Community — real deployment experience
0 Articles · 0 Benchmarks
GPUs that fit (single card)
A100 80GB
80 GB VRAM
$876
/mo · vast
H100 SXM
80 GB VRAM
$1,170
/mo · vast
H200 SXM
141 GB VRAM
$2,555
/mo · lambda
B200
180 GB VRAM
$3,013
/mo · vast
B200 SXM
180 GB VRAM
—