Instructions to use cortexso/phi-4-reasoning with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use cortexso/phi-4-reasoning with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="cortexso/phi-4-reasoning",
	filename="phi-4-mini-reasoning-q2_k.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use cortexso/phi-4-reasoning with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf cortexso/phi-4-reasoning:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf cortexso/phi-4-reasoning:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf cortexso/phi-4-reasoning:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf cortexso/phi-4-reasoning:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf cortexso/phi-4-reasoning:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf cortexso/phi-4-reasoning:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf cortexso/phi-4-reasoning:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf cortexso/phi-4-reasoning:Q4_K_M

Use Docker

docker model run hf.co/cortexso/phi-4-reasoning:Q4_K_M

LM Studio
Jan

vLLM

How to use cortexso/phi-4-reasoning with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "cortexso/phi-4-reasoning"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "cortexso/phi-4-reasoning",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/cortexso/phi-4-reasoning:Q4_K_M

Ollama
How to use cortexso/phi-4-reasoning with Ollama:
```
ollama run hf.co/cortexso/phi-4-reasoning:Q4_K_M
```

Unsloth Studio new

How to use cortexso/phi-4-reasoning with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for cortexso/phi-4-reasoning to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for cortexso/phi-4-reasoning to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for cortexso/phi-4-reasoning to start chatting

Pi new

How to use cortexso/phi-4-reasoning with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf cortexso/phi-4-reasoning:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "phi-4-reasoning"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use cortexso/phi-4-reasoning with Docker Model Runner:
```
docker model run hf.co/cortexso/phi-4-reasoning:Q4_K_M
```

Lemonade

How to use cortexso/phi-4-reasoning with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull cortexso/phi-4-reasoning:Q4_K_M

Run and chat with the model

lemonade run user.phi-4-reasoning-Q4_K_M

List all available models

lemonade list

Overview

Microsoft Research developed and released the Phi-4-reasoning series, a cutting-edge family of reasoning-focused language models optimized for chain-of-thought (CoT), step-by-step problem solving, and high-efficiency inference. These models excel in advanced mathematical reasoning, scientific Q&A, and instruction-following scenarios.

The Phi-4 models introduce extended context lengths, ChatML reasoning templates, and strong performance on benchmark datasets, while maintaining compact sizes that are ideal for memory- and latency-constrained environments.

Variants

Phi-4-reasoning

No	Variant	Branch	Cortex CLI command
1	phi-4-mini-reasoning	4b	`cortex run phi4:4b`
2	phi-4-reasoning	14b	`cortex run phi4:14b`
3	phi-4-reasoning-plus	14b-plus	`cortex run phi4:14b-plus`

Each branch supports multiple quantized GGUF versions:

phi-4-mini-reasoning: q2_k, q3_k_l, q3_k_m, q3_k_s, q4_k_m, q4_k_s, q5_k_m, q5_k_s, q6_k, q8_0
phi-4-reasoning: q2_k, q3_k_l, q3_k_m, q3_k_s, q4_k_m, q4_k_s, q5_k_m, q5_k_s, q6_k, q8_0
phi-4-reasoning-plus: q2_k, q3_k_l, q3_k_m, q3_k_s, q4_k_m, q4_k_s, q5_k_m, q5_k_s, q6_k, q8_0

Use it with Jan (UI)

Install Jan using Quickstart
Use in Jan model Hub:
```
cortexso/phi4
```

Use it with Cortex (CLI)

Install Cortex using Quickstart
Run the model with command:
```
cortex run phi4
```

Credits

Author: Microsoft Research
Converter: Menlo Research
Original License: MIT License
Blogs: Phi-4 Reasoning Blog

Downloads last month: 648

GGUF

Model size

4B params

Architecture

phi3

Hardware compatibility

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit