MicrosoftactiveOpen Source

Phi-4-Mini

phi-4-mini

Compact 3.8B model with 128K context. Strong on-device performance.

Context Window

128K

tokens

Max Output

4.1K

tokens

Input Price

—

per 1M tokens

Output Price

—

per 1M tokens

Details

Familyphi

Parameters3.8B

Training Cutoff2024-09-01

ReleasedFebruary 27, 2025

Capabilities

FunctionsStreamingCodeReasoning

Documentation

Evaluation Scores(3 benchmarks)

HumanEvalFunction-level Python code generation

55%

MATH-500Competition-style math

42%

MMLU-ProHarder successor to MMLU

38%

Quick Access

curl pikaainews.com/api/models/microsoft-phi-4-mini

npx pika-models info microsoft-phi-4-mini

Get API Access

Third-Party Providers & Aggregators

DeepInfra

Lowest per-token rates for open-source models.

Fireworks AI

Fastest inference engine. Multimodal support, HIPAA/SOC2.

Together AI

Fast open-source model inference. Sub-100ms latency.

Other phi models

Microsoft

Phi-4-Reasoning

phi-4-reasoning

32.8K ctx

Microsoft

Phi-4

phi-4

16.4K ctx

Microsoft

Phi-3.5 MoE

phi-3.5-moe

131.1K ctx

Microsoft

Phi-3.5 Mini

phi-3.5-mini

131.1K ctx

Microsoft

Phi-3 Medium

phi-3-medium

131.1K ctx

Microsoft

Phi-3 Mini

phi-3-mini

128K ctx