---
title: "Qwen 2.5 3B Benchmarks - RankedAGI"
description: "Detailed benchmark and metadata record for Qwen 2.5."
source: "https://rankedagi.com/models/qwen-2.5-3b"
---

# Qwen 2.5

| Field | Value |
| --- | --- |
| Organization | Alibaba |
| License | Open Source |
| Size | 3B |
| Version | Latest |
| Released | 2024-09-19 |
| Context window | 32K |
| Knowledge cutoff | 2024-09-01 |
| Last updated | 2026-05-05T14:36:05.509Z |

## Benchmarks

| Benchmark | Category | Value | Description | Source |
| --- | --- | --- | --- | --- |
| RankedAGI Coding | coding | 27.0% | RankedAGI Coding Score |  |
| RankedAGI Agentic | Agents | 31.8% | RankedAGI Agentic Score |  |
| LiveCodeBench v5 | coding | 10.8% | LiveCodeBench - Evaluation of Large Language Models for Code |  |
| RankedAGI Reasoning | reasoning | 27.0% | RankedAGI Reasoning Score |  |
| HumanEval | coding | 84.1% | HumanEval coding benchmark |  |
| HumanEval+ | coding | 75.2% | HumanEval+ benchmark |  |
| MMLU Pro | knowledge | 43.7% | Graduate-level visual problem-solving |  |
| Hughes Hallucination Evaluation | safety | 7.0% | 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 |
| RankedAGI Math | math | 44.9% | RankedAGI Math Score |  |
| RankedAGI Overall | general | 35.8% | Overall RankedAGI score |  |

## Model Links

| Title | URL |
| --- | --- |
| Release | https://qwen.ai/blog?id=qwen2.5 |
| HF | https://huggingface.co/Qwen/Qwen2.5-3B |
| Ollama | https://ollama.com/library/qwen2.5-coder:3b |