Phi-2 2.7B
microsoft/phi-2
Parameters
2.7B
Context
2K
Downloads
848.8K
Likes
3.5K
Architecture
transformer
GPU Requirements
VRAM / RAM / disk estimates by quantization and usage scenario
| Quantization | Minimum | Recommended | Production |
|---|---|---|---|
| FP1616bit | 11.3 GB RAM 32.0 · Disk 46.9 | 27.4 GB RAM 41.0 · Disk 72.9 | 124.5 GB RAM 186.8 · Disk 111.9 |
| FP88bit | 8.1 GB RAM 32.0 · Disk 42.9 | 24.0 GB RAM 36.0 · Disk 68.9 | 121.1 GB RAM 181.6 · Disk 107.9 |
| INT88bit | 8.1 GB RAM 32.0 · Disk 42.9 | 24.0 GB RAM 36.0 · Disk 68.9 | 121.1 GB RAM 181.6 · Disk 107.9 |
| AWQ4bit | 6.6 GB RAM 32.0 · Disk 41.0 | 22.4 GB RAM 33.6 · Disk 67.0 | 119.4 GB RAM 179.1 · Disk 106.0 |
| GPTQ4bit | 6.6 GB RAM 32.0 · Disk 41.1 | 22.5 GB RAM 33.7 · Disk 67.1 | 119.4 GB RAM 179.1 · Disk 106.1 |
| GGUF4bit | 6.7 GB RAM 32.0 · Disk 41.1 | 22.5 GB RAM 33.7 · Disk 67.1 | 119.5 GB RAM 179.2 · Disk 106.1 |
GPU Recommendations
1× RTX 5090
32 GB VRAM
≈ 1378.5 tok/s
Effective production throughput ≈ 964.9 tok/s
score 0.6836
2× Tesla V100
32 GB VRAM
≈ 0 tok/s (N/A — benchmark unavailable)
Effective production throughput ≈ 0 tok/s
score 0.6335
8× H200 SXM
1128 GB VRAM
≈ 11815.4 tok/s
Effective production throughput ≈ 8270.8 tok/s
score 0.2197
1× RTX 3090
24 GB VRAM
≈ 720 tok/s
Effective production throughput ≈ 504 tok/s
score 0.6456
💬 Community — real deployment experience
0 Articles · 0 Benchmarks
GPUs that fit (single card)
RTX 3090
24 GB VRAM
$88.4
/mo · vast
A10
24 GB VRAM
$171
/mo · vast
L4
24 GB VRAM
$185
/mo · vast
RTX 4090
24 GB VRAM
$190
/mo · tensordock
A100 SXM4
80 GB VRAM
$215
/mo · vast