← Reference library
PaperHigh credibilityarXiv:2412.15115 · Qwen Team, Alibaba · December 19, 2024

Qwen2.5 Technical Report

Our summary

The technical report for Qwen2.5, Alibaba's open-weight model family (0.5B–72B) trained on ~18T tokens, with the flagship 72B-Instruct competitive with far larger open models and strong multilingual and long-context support.

Why it matters

Documents a leading open-weight family now in the catalog, and the training scale behind its capabilities and limits.

Published June 26, 2026

Cite this

Qlarify Labs. (2026). Qwen2.5 Technical Report. Retrieved from https://labs.qlarify.fi/references/qwen2.5-technical-report-2024