---
title: "QwQ Preview 32B Benchmarks - RankedAGI"
description: "Detailed benchmark and metadata record for QwQ Preview."
source: "https://rankedagi.com/models/qwq-preview-32b"
---

# QwQ Preview

| Field | Value |
| --- | --- |
| Organization | Alibaba |
| License | Open Source |
| Size | 32B |
| Version | Preview |
| Released | 2024-11-28 |
| Context window | 32K |
| Knowledge cutoff | 2024-09-01 |
| Last updated | 2025-04-16 22:14:46.146909+00 |

## Benchmarks

| Benchmark | Category | Value | Description | Source |
| --- | --- | --- | --- | --- |
| RankedAGI Coding | coding | 33.3% | RankedAGI Coding Score |  |
| RankedAGI Agentic | Agents | 20.6% | RankedAGI Agentic Score |  |
| LiveCodeBench v5 | coding | 44.0% | LiveCodeBench - Evaluation of Large Language Models for Code |  |
| LiveBench Coding Score (old) | coding | 37.2% | LiveBench Coding Score old version |  |
| RankedAGI Reasoning | reasoning | 41.8% | RankedAGI Reasoning Score |  |
| GPQA Diamond | reasoning | 65.2% | Generalized Prefix Question Answering Score (Reasoning) PhD Level Reasoning |  |
| AIME 2025 | math | 33.5% | AIME 2025 Competition Math |  |
| AIME 2024 | math | 44.0% | AIME 2024 Competition Math |  |
| MMLU Pro | knowledge | 75.0% | Graduate-level visual problem-solving |  |
| Hughes Hallucination Evaluation | safety | 12.9% | 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 |
| LiveBench Average | general | 28.5% | LiveBench Average Score old version |  |
| LiveBench Data Analysis | general | 3.2% | LiveBench Data Analysis Score | https://livebench.ai/ |
| LiveBench Language | general | 19.6% | LiveBench Language Score | https://livebench.ai/ |
| RankedAGI Math | math | 47.7% | RankedAGI Math Score |  |
| RankedAGI Overall | general | 38.2% | Overall RankedAGI score |  |