Phi-4 (14B) — VRAM Requirements
作者： Jakub Rusinowski · 最后更新： 2026年7月21日
How much GPU VRAM you need to run Phi-4 Family Phi-4 (14B) by Microsoft locally, a 14B-parameter model. Figures are quantized weights + KV cache + framework overhead from a memory-bandwidth roofline model.
Phi-4 (14B) needs about 11 GB VRAM at Q4_K_M.
VRAM by Quantization
| Quant | Bits/weight | Weights | Total VRAM |
|---|---|---|---|
| Q2_K | 2.63 | 4.6 GB | 7.1 GB |
| Q3_K_M | 3.41 | 6.0 GB | 8.4 GB |
| Q4_K_M | 4.83 | 8.5 GB | 10.9 GB |
| Q5_K_M | 5.67 | 9.9 GB | 12.4 GB |
| Q6_K | 6.56 | 11.5 GB | 14.0 GB |
| Q8_0 | 8.50 | 14.9 GB | 17.4 GB |
| F16 | 16.00 | 28.0 GB | 30.5 GB |
Switch quantization in the interactive calculator, or see the full Phi-4 Family model page.
Add this badge to your model card
Model creators: paste this into your Hugging Face model card README to link readers straight to this VRAM breakdown.
[](https://llmconfigurator.com/en/vram-calculator/phi-4-14b)Estimates only — actual VRAM varies with context length, batch size, runtime and KV-cache settings.