Run Phi 4 locally: hardware requirements
Phi 4 by Microsoft installs on a paired computer through idapt's local engine. RAM decides whether a build fits; a GPU only changes speed. Local runs are not billed.
- Download
- 9.1 GB-16 GB
- RAM needed
- 12 GB-20 GB
- Builds
- 2
Will it run on your machine?
Enter your system RAM to get verdicts. RAM decides whether a build fits; a GPU only changes speed.
Every build
| Build | Quantization | Download | RAM needed |
|---|---|---|---|
| 14b-q4_K_MDefault | q4_K_M | 9.1 GB | 12 GB |
| 14b-q8_0 | q8_0 | 16 GB | 20 GB |
RAM figures are working-set recommendations: when GPU memory is short, the runtime spills to CPU and system memory instead of failing.
Install it
Pair a computer with idapt and install Phi 4 from the model page: the local engine bootstraps itself, pulls the build you pick, and streams progress. Or pull it directly:
ollama pull phi4:14b-q4_K_MFrequently asked
How much RAM do I need to run Phi 4 locally?
The smallest Phi 4 build runs with about 12 GB of system RAM; the largest needs 20 GB. RAM is the binding requirement: when GPU memory is short, the runtime spills to CPU and system memory instead of failing.
Which quantization of Phi 4 should I pick?
Start with the default 14b-q4_K_M build: a 9.1 GB download that wants about 12 GB of RAM. Smaller quantizations shrink the download and memory needs at some cost in output quality; larger ones do the reverse.
Do I need a GPU to run Phi 4?
No. A GPU with enough VRAM makes generation faster, but Phi 4 runs on CPU and system RAM alone. Whether a build fits is decided by RAM; the GPU only changes speed.
How big is the Phi 4 download?
Between 9.1 GB and 16 GB, depending on the quantization you pick across 2 available builds.
Can idapt install Phi 4 for me?
Yes. Pair a computer with idapt and install Phi 4 from this page: the local engine bootstraps itself, pulls the build you pick (ollama pull phi4), and streams download progress. Local runs are not billed.
Part of the Phi 4 model page · Phi 4 pricing · Check every local model against your hardware