Phi 4 Mini

Phi 4 Mini by Microsoft. A small multilingual model.

Added Jul 26, 2025

Pricing

Auto routing · per 1M tokens

Input: $0.17
Output: $0.68

Compare provider prices

Specifications

Context window: 128K
Max output: 16.4K
Parameters: 3.8B
Avg output (7d): 38 tokens; Longer than 2% of models

Benchmarks

Artificial Analysis

LMArena

Vectara

Sourced from Artificial Analysis.

Intelligence Index

6.3

Better than 15% of models compared

Coding Index

3.8

Better than 3% of models compared

Agentic work

T²-Bench Telecom (legacy)

Legacy fallback · Conversational AI agents in dual-control scenarios

8.2%

Better than 6% of models compared

GDPval-AA v2

Economically valuable tasks (Elo)

-405 Elo

Better than 0% of models compared

Document reasoning

AA-LCR v1.1

Long context reasoning with updated grading

15.3%

Better than 18% of models compared

Reasoning

HLE

Humanity's Last Exam

4.4%

Better than 20% of models compared

IFBench

Instruction-following benchmark

21.1%

Better than 3% of models compared

CritPt

Research-level physics reasoning

0.0%

Coding

Terminal-Bench Hard (legacy)

Legacy fallback · Agentic coding and terminal use

0.0%

Better than 5% of models compared

LiveCodeBench

Contamination-free coding benchmark

12.6%

Better than 12% of models compared

Math

AIME 2025

American Invitational Mathematics Examination 2025

6.7%

Better than 10% of models compared

AIME

American Invitational Mathematics Examination

3.0%

Better than 14% of models compared

Math-500

Diverse mathematical problem solving benchmark

69.6%

Better than 28% of models compared

Knowledge

MMLU-Pro

Professional and academic subject knowledge

46.5%

Better than 12% of models compared

AA-Omniscience Accuracy

Proportion of correctly answered questions

14.1%

AA-Omniscience Hallucination Rate

Rate of incorrect answers among non-correct responses

81.2%

Legacy benchmarks

GPQA Diamond (legacy)

Graduate-level scientific reasoning

33.1%

Better than 10% of models compared

AA-LCR (unversioned / legacy)

Long context reasoning evaluation