v2.2 · 2026-06-12

Phi-4 Mini Reasoning

phi-4-mini-reasoning

REASONING

Phi-4 Mini Reasoning is a 3.8B reasoning model profile for local AI planning. This page records provider, license, context, quantization, hardware-fit estimates, setup hints, source links, and caveats so readers do not choose a model by name alone.

PROVIDER

Microsoft

FAMILY

Phi-4

MODEL TYPE

reasoning

PARAMETERS

3.8B

MODALITIES

text

ARCHITECTURE

Dense decoder-only Transformer based on Phi-4 Mini; math-reasoning fine-tune

CONTEXT WINDOW

128,000 tokens

TRAINING TOKENS

—

RELEASE DATE

—

LICENSE & LINKS

MITlicense text

MODEL CARD

huggingface.co/microsoft/Phi-4-mini-reasoning

HUGGING FACE

microsoft/Phi-4-mini-reasoning

GGUF REPOSITORIES

bartowski/microsoft_Phi-4-mini-reasoning-GGUF community conversion

Community GGUF repository referenced by the June 2026 model research.

lmstudio-community/Phi-4-mini-reasoning-GGUF community conversion

Community GGUF repository referenced by the June 2026 model research.

SETUP HINTS

OLLAMA

ollama run phi4-mini-reasoning

LM STUDIO

Available according to the supplied June 2026 research; select a cited GGUF/MLX build and verify current app metadata before relying on hands-on setup guidance.

LLAMA.CPP

Use the cited GGUF repository (bartowski/microsoft_Phi-4-mini-reasoning-GGUF) with current llama.cpp-compatible tooling.

QUANTIZATION FILE SIZES (GGUF)

Q4_K_M

2.49 GB

Q5_K_M

2.85 GB

Q8_0

4.08 GB

Quantization note: Q4_K_M for first local attempts unless a cited runtime recommends a different default.

RAM / VRAM ESTIMATES

MIN RAM

8 GB

COMFORTABLE RAM

16 GB

MIN VRAM

8 GB

COMFORTABLE VRAM

16 GB

These are conservative local-inference estimates, not official hardware requirements.

HARDWARE FIT

CPU only (no GPU)

limited

8 GB RAM

limited

16 GB RAM

comfortable

32 GB RAM

comfortable

8 GB VRAM

limited

12 GB VRAM

limited

24 GB VRAM

comfortable

Apple Silicon (unified memory)

comfortable

Hardware fit values are conservative local-inference estimates based on GGUF size plus runtime and KV-cache overhead. Actual requirements depend on context length, quantization, runtime, GPU offload, and other running apps.

BEST FOR

·Math practice, symbolic problem solving, proof-style prompts, and lightweight reasoning experiments in local apps today.

AVOID IF

·Avoid for general knowledge, multilingual chat, image tasks, or polished assistant-style writing and brainstorming work.

CAVEATS

·Microsoft describes this model as designed and tested for math reasoning only. Treat it as a specialty model, not a general assistant. It may be verbose, may over-reason simple prompts, and has limited factual knowledge.

·Commercial-use status from the supplied research: Permitted under MIT license; verify application-specific obligations before commercial deployment.

·Beginner summary from supplied research: Phi-4 Mini Reasoning is the beginner-local math specialty choice: small enough for Ollama, LM Studio, GGUF, and MLX, but tuned specifically for step-by-step mathematical reasoning rather than everyday factual chat or broad writing tasks locally.

SOURCE URLS