Model family

Phi local AI models

Explore the Phi models currently tracked by AIPCFit. Compare parameter sizes, available GGUF variants, catalog use cases, and hardware-fit guidance without treating model size as a universal performance guarantee.

3

Models

3

GGUF variants

3.8B–14B

Parameter range

Microsoft

Provider

Catalog use cases

chatlightweightcodingreasoning

Microsoft

Phi-3.5 Mini

small

3.8B parameters · 1 GGUF variant

chat · lightweight · coding

Open model →3 GPU compatibility checks

Microsoft

Phi-4 Mini

small

3.8B parameters · 1 GGUF variant

chat · lightweight · coding

Open model →3 GPU compatibility checks

Microsoft

Phi-4

medium

14B parameters · 1 GGUF variant

chat · reasoning · coding

Open model →3 GPU compatibility checks

GPU × model

Browse compatibility checks

See published dedicated-VRAM fit checks for selected NVIDIA RTX GPUs and local models.

Your hardware

What LLM can I run?

Start from your actual PC or Mac instead of a model family.