---
title: "Gemma 3 4B Benchmarks - RankedAGI"
description: "Detailed benchmark and metadata record for Gemma 3."
source: "https://rankedagi.com/models/gemma-3-4b"
---

# Gemma 3

| Field | Value |
| --- | --- |
| Organization | Google |
| License | Open Source |
| Size | 4B |
| Version | Latest |
| Released | 2025-03-12 |
| Last updated | 2026-04-27T17:10:20.797Z |

## Benchmarks

| Benchmark | Category | Value | Description | Source |
| --- | --- | --- | --- | --- |
| RankedAGI Coding | coding | 21.6% | RankedAGI Coding Score |  |
| RankedAGI Agentic | Agents | 17.1% | RankedAGI Agentic Score |  |
| LiveCodeBench v6 | coding | 11.0% | LiveCodeBench v6 - Evaluation of Large Language Models for Code |  |
| RankedAGI Reasoning | reasoning | 23.7% | RankedAGI Reasoning Score |  |
| Humanity's Last Exam | reasoning | 5.0% | Multidisciplinary Reasoning (no tools) |  |
| GPQA Diamond | reasoning | 29.0% | Generalized Prefix Question Answering Score (Reasoning) PhD Level Reasoning |  |
| AIME 2025 | math | 13.0% | AIME 2025 Competition Math |  |
| Instruction Following Evaluation | instruction-following | 28.0% | Instruction Following Evaluation |  |
| LiveBench Coding 25.4 | coding | 11.7% | LiveBench Coding Score 2025-04 | https://livebench.ai/ |
| MMMU-Pro | imaging | 30.0% | Multimodal Understanding |  |
| RankedAGI Math | math | 33.6% | RankedAGI Math Score |  |
| RankedAGI Overall | general | 30.3% | Overall RankedAGI score |  |
| 𝜏²-Bench Telecom | Agents | 5.0% | Agentic Tool Use |  |

## Model Links

| Title | URL |
| --- | --- |
| Release | https://blog.google/technology/developers/gemma-3/ |