Written by Jakub Rusinowski · Last updated February 4, 2025
Model library → Phi-4 Mini → Phi-4 Mini (3.8B)
Significant upgrade over Phi-3.5 Mini in reasoning and coding. Tiny footprint makes it ideal for iPhone, Android, and laptops without a GPU. MIT licensed for commercial use.
Phi-4 Mini (3.8B) needs about 3 GB of VRAM at Q4_K_M — quantized weights plus framework overhead, before any KV cache. On Apple Silicon that figure comes out of unified memory.
| Parameters | 3.8 Billion |
| Context window | 128,000 |
| Architecture | Dense |
| Provider | Microsoft |
| Licence | MIT |
| Specified at | Q4_K_M |
| System RAM | 6 GB |
| Record updated | 2025-02-04 |
MIT — commercial use permitted. Commercial use permitted. No usage restrictions beyond attribution.
Modelled on a reference NVIDIA RTX 4090 (24 GB), at 8K context. Speed figures are ESTIMATES from the memory-bandwidth roofline described on the methodology page, not benchmarks we ran — rows marked measured come from published or reader-submitted runs. VRAM here includes the KV cache, so it reads higher than the headline figure above, which does not.
| Quant | Weights | VRAM needed | Est. speed | Fit on 24 GB |
|---|---|---|---|---|
| Q2_K | 1.2 GB | 3.1 GB | ~217 tok/s (est.) | Fits comfortably |
| Q3_K_M | 1.6 GB | 3.5 GB | ~196 tok/s (est.) | Fits comfortably |
| Q4_K_M | 2.3 GB | 4.2 GB | ~167 tok/s (est.) | Fits comfortably |
| Q5_K_M | 2.7 GB | 4.6 GB | ~153 tok/s (est.) | Fits comfortably |
| Q6_K | 3.1 GB | 5.0 GB | ~141 tok/s (est.) | Fits comfortably |
| Q8_0 | 4.0 GB | 5.9 GB | ~121 tok/s (est.) | Fits comfortably |
| F16 | 7.6 GB | 9.5 GB | ~77 tok/s (est.) | Fits comfortably |
Want the memory numbers alone, at every quantization level and your own context length? Use the Phi-4 Mini (3.8B) VRAM calculator.
or compare on Vast.ai from $0.35/hr (typical low · varies)
As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.
The cheapest catalogued GPU that runs Phi-4 Mini (3.8B) is the Intel Arc B570 (10 GB).
Install Ollama, then run:
ollama run phi4-miniWeights on Hugging Face: microsoft/Phi-4-mini-instruct.
Best for: mobile, edge devices, reasoning, fast chat.
← All Phi-4 Mini models | VRAM calculator | Check your own hardware