---
title: "Phi 4 14B Benchmarks - RankedAGI"
description: "Detailed benchmark and metadata record for Phi 4."
source: "https://rankedagi.com/models/phi-4-14b"
---

# Phi 4

| Field | Value |
| --- | --- |
| Organization | Microsoft |
| License | Open Source |
| Size | 14B |
| Version | Latest |
| Released | 2024-12-12 |
| Last updated | 2025-04-16 22:14:46.146909+00 |

## Benchmarks

| Benchmark | Category | Value | Description | Source |
| --- | --- | --- | --- | --- |
| RankedAGI Coding | coding | 32.7% | RankedAGI Coding Score |  |
| RankedAGI Agentic | Agents | 30.5% | RankedAGI Agentic Score |  |
| RankedAGI Reasoning | reasoning | 40.0% | RankedAGI Reasoning Score |  |
| GPQA Diamond | reasoning | 56.1% | Generalized Prefix Question Answering Score (Reasoning) PhD Level Reasoning |  |
| HumanEval | coding | 82.6% | HumanEval coding benchmark |  |
| HumanEval+ | coding | 82.8% | HumanEval+ benchmark |  |
| NYT Connections Extended | reasoning | 10.2% | NYT Connections Extended Version | https://github.com/lechmazur/nyt-connections/ |
| MMLU Pro | knowledge | 70.4% | Graduate-level visual problem-solving |  |
| MMLU | knowledge | 84.8% | Image Reasoning |  |
| Instruction Following Evaluation | instruction-following | 63.0% | Instruction Following Evaluation |  |
| RankedAGI Math | math | 49.8% | RankedAGI Math Score |  |
| RankedAGI Overall | general | 39.9% | Overall RankedAGI score |  |
