TLDRocket
Sign in

Qwen2

Model ● Covered in 6 stories + Follow

Qwen2 is an open-source large language model family developed by Alibaba, with base models ranging from 0.5B to 72B parameters that support 128K token context windows and 29+ languages. Recent releases include Qwen2.5, which expanded pre-training data to 18 trillion tokens and achieved MMLU scores above 85, and Qwen2-VL, a multimodal vision-language variant capable of processing videos longer than 20 minutes. The models have been integrated into Text Generation Inference for deployment on Intel Gaudi hardware and are being used in agent-based systems for extended context processing.

Updated 4 August 2026

Specifications

No specifications recorded yet.

Latest developments

🚀 Accelerating LLM Inference with TGI on Intel Gaudi Hugging Face · 1 year ago · 37
Qwen2.5: A Party of Foundation Models! GitHub Pages · 1 year ago · 48
Qwen2.5-LLM: Extending the boundary of LLMs GitHub Pages · 1 year ago · 20
Qwen2-VL: To See the World More Clearly GitHub Pages · 2 years ago · 47
Hello Qwen2 GitHub Pages · 2 years ago · 53
Generalizing an LLM from 8k to 1M Context using Qwen-Agent GitHub Pages · 2 years ago · 40

Timeline

Month Quarter Year

March 2025

September 2024

Alibaba releases Qwen2.5 family of open-source language models with coding and math variants Open source release

August 2024

June 2024

Relationships

Products & technology

  • Alibaba develops this model · 2 sources
  • Derived from Qwen1.5 · 1 source
  • Qwen2.5 derived from this model · 1 source
  • Qwen2-VL derived from this model · 1 source

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.