Phi-3.5 Mini 3.8B
microsoft/Phi-3.5-mini-instruct
3.8B mini coding-capable model, runs anywhere.
Parameters
3.8B
Context
128K
Downloads
610K
Likes
1.2K
Architecture
transformer
GPU Requirements
VRAM / RAM / disk estimates by quantization and usage scenario
| Quantization | Minimum | Recommended | Production |
|---|---|---|---|
| FP1616bit | 14.0 GB RAM 32.0 · Disk 50.1 | 30.1 GB RAM 45.2 · Disk 76.1 | 127.4 GB RAM 191.1 · Disk 115.1 |
| FP88bit | 9.5 GB RAM 32.0 · Disk 44.6 | 25.4 GB RAM 38.1 · Disk 70.6 | 122.5 GB RAM 183.8 · Disk 109.6 |
| INT88bit | 9.5 GB RAM 32.0 · Disk 44.6 | 25.4 GB RAM 38.1 · Disk 70.6 | 122.5 GB RAM 183.8 · Disk 109.6 |
| AWQ4bit | 7.3 GB RAM 32.0 · Disk 41.9 | 23.1 GB RAM 34.7 · Disk 67.9 | 120.2 GB RAM 180.2 · Disk 106.9 |
| GPTQ4bit | 7.3 GB RAM 32.0 · Disk 41.9 | 23.2 GB RAM 34.8 · Disk 67.9 | 120.2 GB RAM 180.3 · Disk 106.9 |
| GGUF4bit | 7.4 GB RAM 32.0 · Disk 42.0 | 23.2 GB RAM 34.8 · Disk 68.0 | 120.2 GB RAM 180.4 · Disk 107.0 |
GPU Recommendations
1× RTX 5090
32 GB VRAM
≈ 968.6 tok/s
Effective production throughput ≈ 678 tok/s
score 0.6906
2× Tesla V100
32 GB VRAM
≈ 0 tok/s (N/A — benchmark unavailable)
Effective production throughput ≈ 0 tok/s
score 0.6399
8× H200 SXM
1128 GB VRAM
≈ 8302.7 tok/s
Effective production throughput ≈ 5811.9 tok/s
score 0.2197
1× RTX 3090
24 GB VRAM
≈ 505.9 tok/s
Effective production throughput ≈ 354.1 tok/s
score 0.6417
💬 Community — real deployment experience
0 Articles · 0 Benchmarks
GPUs that fit (single card)
RTX 3090
24 GB VRAM
$78.7
/mo · vast
RTX 4090
24 GB VRAM
$98.1
/mo · vast
RTX 4090D
24 GB VRAM
$117
/mo · vast
L4
24 GB VRAM
$148
/mo · vast
A10
24 GB VRAM
$176
/mo · vast