Qwen

Qwen is the large language model family developed by Alibaba Cloud, with most variants released under the Apache 2.0 license. The lineup spans dense and mixture-of-experts models for general chat, plus specialized lines for coding and vision-language tasks, and the open weights have made Qwen one of the most downloaded and fine-tuned model families on Hugging Face. Hosted access runs through Alibaba Cloud's Model Studio, while the weights self-host through standard runtimes. Frequent releases and permissive licensing have made the family a common base for fine-tuned derivatives across the open community. Variants span small models that run on a laptop up to large mixture-of-experts systems for demanding tasks.

Key Capabilities:

• Open-weight dense and mixture-of-experts models under Apache 2.0

• Qwen-Coder models for code generation and repository tasks

• Vision-language models for multimodal understanding

• OpenAI-compatible API through Alibaba Cloud Model Studio

• Long-context variants for large-document processing

• Self-hostable through vLLM, SGLang, and Ollama