Autor: Jakub Rusinowski · Ostatnia aktualizacja: 4 lutego 2025
Ultra-wydajny mały model Microsoftu do wdrożeń mobilnych i edge. Phi-4 Mini osiąga niezwykłe możliwości rozumowania w zaledwie 3,8B parametrach, zaprojektowany do AI na urządzeniu na telefonach i laptopach bez dedykowanych GPU.
| Licence | What it permits | Applies to |
|---|---|---|
MIT | Commercial use permitted Commercial use permitted. No usage restrictions beyond attribution. | Phi-4 Mini (3.8B) |
| Phi-4 Mini (3.8B) | Min 3 GB VRAM · Q4_K_M · 128,000 ctx · ollama run phi4-mini |
The cheapest GPU that runs Phi-4 Mini locally (min 3 GB VRAM) is the Intel Arc B570 (10 GB).
Install Ollama then run: ollama run phi4-mini
Minimum VRAM: 3 GB. For best results use Q4_K_M quantization.
Phi-4 Mini needs about 3 GB VRAM at Q4_K_M quantization for its smallest variant. Variants: Phi-4 Mini (3.8B) (3 GB, Q4_K_M). On Apple Silicon, unified memory counts toward this requirement.
Yes — Phi-4 Mini runs on an RTX 4090 (24 GB) and other 24 GB cards such as the RTX 3090. Smaller variants also fit comfortably on 8–16 GB GPUs at Q4_K_M.
Q4_K_M is the best balance of quality and VRAM for Phi-4 Mini in most cases. Choose Q8_0 for near-lossless quality if you have spare VRAM, or smaller quants (Q3/Q2) only when memory is tight.
Install Ollama, then run: ollama run phi4-mini. This downloads Phi-4 Mini and starts a local, OpenAI-compatible endpoint — no internet connection is needed after the initial download.