Qwen2

Qwen2 is an advanced suite of foundational and instruction-tuned language models, with parameters ranging from 0.5 to 72 billion. It includes both dense models and a Mixture-of-Experts model.

Qwen2 outperforms most previous open-weight models, including its predecessor Qwen1.5, and demonstrates competitive performance against proprietary models across various benchmarks, including language understanding, generation, multilingual capabilities, coding, mathematics, and reasoning.

Qwen Version 2 showcases strong multilingual abilities, excelling in around 30 languages, including English, Chinese, Spanish, French, German, Arabic, Russian, Korean, Japanese, Thai, Vietnamese, and more, highlighting its versatility and global reach.

We are an independent community dedicated to exploring and understanding the Qwen2 models.