---
title: "Gemma 2 9B Benchmarks - RankedAGI"
description: "Detailed benchmark and metadata record for Gemma 2."
source: "https://rankedagi.com/models/gemma-2-9b"
---

# Gemma 2

| Field | Value |
| --- | --- |
| Organization | Google |
| License | Open Source |
| Size | 9B |
| Version | Latest |
| Released | 2024-06-27 |
| Context window | 8K |
| Knowledge cutoff | 2024-06-01 |
| Last updated | 2025-04-16 22:14:46.146909+00 |

## Benchmarks

| Benchmark | Category | Value | Description | Source |
| --- | --- | --- | --- | --- |
| RankedAGI Coding | coding | 26.9% | RankedAGI Coding Score |  |
| RankedAGI Agentic | Agents | 33.8% | RankedAGI Agentic Score |  |
| ChatArena (LMSYS) | coding | 1180 | ChatArena (LMSYS) Coding ELO Score with Style Control |  |
| LiveBench Coding Score (old) | coding | 22.5% | LiveBench Coding Score old version |  |
| RankedAGI Reasoning | reasoning | 29.5% | RankedAGI Reasoning Score |  |
| HumanEval | coding | 40.2% | HumanEval coding benchmark |  |
| MMLU Pro | knowledge | 52.1% | Graduate-level visual problem-solving |  |
| MMLU | knowledge | 71.3% | Image Reasoning |  |
| Hughes Hallucination Evaluation | safety | 10.1% | Computed using Vectara's Hallucination Evaluation Model, also known as HHEM. This evaluates how often an LLM introduces hallucinations when summarizing a document. | https://github.com/vectara/hallucination-leaderboard |
| AidanBench | general | 859 | AidanBench Score |  |
| LiveBench Average | general | 28.4% | LiveBench Average Score old version |  |
| LiveBench Data Analysis | general | 35.1% | LiveBench Data Analysis Score | https://livebench.ai/ |
| LiveBench Language | general | 25.5% | LiveBench Language Score | https://livebench.ai/ |
| RankedAGI Math | math | 50.8% | RankedAGI Math Score |  |
| RankedAGI Overall | general | 37.0% | Overall RankedAGI score |  |