// npm install @heyputer/puter.js
import { puter } from '@heyputer/puter.js';
puter.ai.chat("Explain quantum computing in simple terms", {
model: "microsoft/phi-4"
}).then(response => {
document.body.innerHTML = response.message.content;
});
<html>
<body>
<script src="https://js.puter.com/v2/"></script>
<script>
puter.ai.chat("Explain quantum computing in simple terms", {
model: "microsoft/phi-4"
}).then(response => {
document.body.innerHTML = response.message.content;
});
</script>
</body>
</html>
# pip install openai
from openai import OpenAI
client = OpenAI(
base_url="https://api.puter.com/puterai/openai/v1/",
api_key="YOUR_PUTER_AUTH_TOKEN",
)
response = client.chat.completions.create(
model="microsoft/phi-4",
messages=[
{"role": "user", "content": "Explain quantum computing in simple terms"}
],
)
print(response.choices[0].message.content)
curl https://api.puter.com/puterai/openai/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_PUTER_AUTH_TOKEN" \
-d '{
"model": "microsoft/phi-4",
"messages": [
{"role": "user", "content": "Explain quantum computing in simple terms"}
]
}'
Model Card
Phi-4 is a 14B parameter small language model from Microsoft that excels at complex reasoning tasks, especially mathematics, outperforming many larger models on math competition benchmarks while being efficient enough for edge deployment.
Context Window 16K
tokens
Max Output 16K
tokens
Input Cost $0.07
per million tokens
Output Cost $0.14
per million tokens
Release Date Jan 10, 2025
Output Speed 38
tokens / sec
Latency 0.51s
time to first token
Model Playground
Try Phi 4 instantly in your browser.
This playground uses the Puter.js AI API — no API keys or setup required.
Benchmarks
How Phi 4 performs on standard evaluations.
| Benchmark | Score |
|---|---|
| GPQA Diamond Graduate-level science Q&A | 57.5% |
| Humanity's Last Exam Cross-domain reasoning | 4.1% |
| LiveCodeBench Recent coding problems | 23.1% |
| SciCode Scientific programming | 26.0% |
| MATH-500 Competition math | 81.0% |
| AIME 2024 Advanced math exam | 14.3% |
| AIME 2025 Advanced math exam | 18.0% |
| IFBench Instruction following | 23.5% |
| LCR Long-context reasoning | 0.0% |
| Terminal-Bench Hard Agentic terminal tasks | 3.8% |
| τ²-Bench Tool use / agents | 0.0% |
Scores sourced from Artificial Analysis.
Find other Microsoft models →
Phi 4 Mini
Phi-4 Mini is a 3.8-billion-parameter small language model developed by Microsoft, designed to deliver strong reasoning performance in a compact form factor. It uses a dense decoder-only Transformer architecture with grouped-query attention and supports context lengths up to 128K tokens. The model excels at math, logic, coding, instruction following, and function calling — making it well-suited for agentic workflows that integrate external tools and APIs. It supports over 20 languages thanks to its expanded 200K-token vocabulary. Despite its small size, Phi-4 Mini performs competitively with much larger models on text-based reasoning tasks. It scored 88.6% on GSM8K and 83.7% on ARC-Challenge. It's a strong choice for developers who need capable reasoning at low latency and minimal compute cost.
ChatWizardLM-2 8x22B
WizardLM-2 8x22B is Microsoft AI's most advanced Wizard model, a Mixture of Experts LLM fine-tuned from Mixtral 8x22B that delivers near-GPT-4 performance on complex chat, multilingual, reasoning, and coding tasks while remaining open-source.
Frequently Asked Questions
You can access Phi 4 by Microsoft through Puter.js AI API. Include the library in your web app or Node.js project and start making calls with just a few lines of JavaScript — no backend and no configuration required. You can also use it with Python or cURL via Puter's OpenAI-compatible API.
Yes, it is free if you're using it through Puter.js. With the User-Pays Model, you can add Phi 4 to your app at no cost — your users pay for their own AI usage directly, making it completely free for you as a developer.
| Price per 1M tokens | |
|---|---|
| Input | $0.07 |
| Output | $0.14 |
Phi 4 was created by Microsoft and released on Jan 10, 2025.
Phi 4 supports a context window of 16K tokens. For reference, that is roughly equivalent to 33 pages of text.
Phi 4 can generate up to 16K tokens in a single response.
Phi 4 scores 10.4 on the Artificial Analysis Intelligence Index, outperforming 16% of tracked models. On coding, it scores 11.2 (outperforms 27% of models). On math, it scores 18.0 (outperforms 20% of models).
Yes — the Phi 4 API works with any JavaScript framework, Node.js, or plain HTML through Puter.js. Just include the library and start building. See the documentation for more details.
Get started with Puter.js
Add Phi 4 to your app without worrying about API keys or setup.
Read the Docs View Tutorials