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

Libraries

How to use microsoft/phi-4 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="microsoft/phi-4")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("microsoft/phi-4")
model = AutoModelForCausalLM.from_pretrained("microsoft/phi-4")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use microsoft/phi-4 with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/microsoft/phi-4

SGLang

How to use microsoft/phi-4 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "microsoft/phi-4" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "microsoft/phi-4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "microsoft/phi-4" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "microsoft/phi-4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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

Phi-4 Training Cutoff Date (Responds October 2023)

#45

by kittikitti - opened Apr 20, 2025

Discussion

kittikitti

Apr 20, 2025

Overview

On the model card we see here,

Dates October 2024 – November 2024

Also, Microsoft's official documentation at https://azure.microsoft.com/en-us/products/phi states in their FAQ,

However, when prompting the model responds October 2023.

Example 1

user: What is the cutoff date for Phi training data?
phi-4: As a large language model, my training data includes information up until October 2023. This means I can provide information and insights based on developments and knowledge available up to that point. It's important to stay updated with the latest sources for the most current developments beyond this date.

Example 2

user: When is your training cutoff date?
phi-4: As of my last update, my training data includes information up to October 2023. This means I can provide information and insights based on knowledge available up to that date. If you have questions about developments or events beyond this time, I recommend consulting more current resources. How else may I assist you today?

Discussion

Is this intended behavior?
Is this a hallucination?
Which prompts illicit a correct response?
If it's a hallucination or otherwise incorrect, can this be fixed?

gugarosa

Microsoft org May 9, 2025

•

edited May 9, 2025

Hello @kittikitti !

According to the model card, "Dates" are referred to the period when the model has been trained, whereas the "Status" indicates the model's knowledge cutoff. It is expected that for publicly available data the model answers with a date earlier than June 2024.

kittikitti

May 18, 2025

Hello @kittikitti !

According to the model card, "Dates" are referred to the period when the model has been trained, whereas the "Status" indicates the model's knowledge cutoff. It is expected that for publicly available data the model answers with a date earlier than June 2024.

Thank you, this is a good explanation.

gugarosa changed discussion status to closed Jul 8, 2025

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment