How much does Phi-4 cost?

Phi-4 costs $0.07 per million input tokens and $0.14 per million output tokens on standard pricing. A workload of 20M input and 5M output tokens a month is about $2.10.

What is the context length of Phi-4?

Phi-4 accepts up to 16,384 tokens (16K) of input and can return up to 14,745 tokens in one response.

Does Phi-4 support tool calling and structured outputs?

Tool calling is not listed for Phi-4 and structured (JSON) outputs. Supported tools: Structured outputs, JSON mode, Deterministic seed.

What inputs and outputs does Phi-4 support?

Phi-4 accepts Text as input and produces Text as output.

Which providers serve Phi-4?

Phi-4 is available from 1 API provider: DeepInfra. Prices can differ by provider.

When was Phi-4 released?

Phi-4 was released by Microsoft in 2024-12.

Phi-4

phi-4

API Provider site

14B small language model, great for edge and cheap tasks. Built by Microsoft.

Prices updated recently

Input price

$0.07

per 1M tokens · Standard

Output price

$0.14

per 1M tokens · Standard

Input limit

16K

tokens

Output limit

15K

tokens

Input formats

TextImageVideoAudioPDF

Output formats

TextImageVideoAudioPDF

Compare

Phi-4 price history

12 price records since 4/1/2026 · 10 changes

Phi-4 cost calculator

Input20M tokens

Output5M tokens

$2.10/month

Standard pricing

GPT-4o mini$6.00

Overview

What is Phi-4?

Phi-4 is a code & slms and text & reasoning model from Microsoft. 14B small language model, great for edge and cheap tasks. Its 16K context window and $0.07 input price make it a candidate for cost-sensitive, high-throughput applications.

Code generationAgent workflowsContent workflowsClassification

Benchmarks

Phi-4 benchmarks & speed

How Phi-4 scores on standardized evaluations, and where it lands among every model we track.

5.9

Intelligence Index

Better than 5% of 153 models

39tok/s

Output speed

Better than 4% of 119 models

Reasoning

GPQA Diamond

57.5%

Graduate-level scientific reasoning · top 82%

HLE

3.8%

Humanity's Last Exam · top 92%

Latency & design

Time to first token: 2.59s

Independent scores from Artificial Analysis and DesignArena · updated 10/5/2026. Higher is better; ranks compare against every model we track with that score.

Rates

Phi-4 pricing

Pricing is based on token usage. Provider-specific caching, batch, regional, and tool charges may affect the final cost.

Input · Standard

$0.07

Per 1M tokens

Cached input · Standard

Not available

No listed cached-input rate

Capabilities

Phi-4 Tools

Tools available when using Phi-4 through supported provider APIs.

Function calling

Not supported

Structured outputs

Supported

JSON mode

Supported

Reasoning

Not supported

Built-in web search

Not supported

Log probabilities

Not supported

Deterministic seed

Supported

Parallel tool calls

Not supported

Prompt caching

Not supported

Strengths and limitations

Strengths

Low input cost at $0.07 per million tokens suits high-volume workloads.
A 16K context window covers most focused application workflows.
Supports code & slms and text & reasoning workloads in one model.

Limitations

Real-world cost still depends on prompt length, response length and provider-specific billing rules.
The 16K context window is smaller than several long-context alternatives.
Arena Elo and MMLU-Pro are directional; test accuracy, latency and reliability on your own workload before committing.

Arena Elo, MMLU-Pro and pricing figures are illustrative; validate current vendor terms before purchase.

Same provider