← All modelsMODEL CHECK
Can I run Phi-4 Mini 3.8B?
Phi-4 Mini 3.8B by Microsoft needs around 6 GB of RAM at the recommended 4-bit quantization (2.3 GB download). Your hardware is checked below — instantly, nothing leaves your browser.
Reading your hardware signals…
Specifications
Parameters3.8B
Context window128K tokens
ProviderMicrosoft
LicenseMIT
Released2025-02
Best forChat
Size by quantization
| Quantization | Bits/weight | Download | Min RAM | Quality |
|---|---|---|---|---|
| Q2_K | 3.35 | 1.6 GB | 4 GB | Noticeable loss |
| Q4_K_MRecommended | 4.85 | 2.3 GB | 6 GB | Recommended |
| Q5_K_M | 5.65 | 2.7 GB | 6 GB | High |
| Q8_0 | 8.5 | 4.0 GB | 8 GB | Near-original |
| F16 | 16 | 7.6 GB | 12 GB | Original |
Sizes are estimates from parameter count × bits per weight; real GGUF builds vary slightly.
Run it locally
The easiest path is Ollama — one command and you're chatting:
ollama run phi4-miniSources & downloads
Ollama Library
Pull and run the model with a single command.
ollama.comHugging Face
Model weights, files and license details.
huggingface.coOfficial GitHub repo
Source code, releases and issues from Microsoft.
github.comMicrosoft — official page
Official page and documentation from Microsoft.
azure.microsoft.com