Qwen is Alibaba's open-weight LLM family and the most downloaded open model family globally in 2026. The Qwen3 series spans sizes from 0.5B to 72B parameters, with specialized variants for coding, math, and vision. Qwen is particularly strong for multilingual applications, especially Chinese, Japanese, Korean, and other Asian languages where Western models underperform.
Qwen models are released on Hugging Face under Apache 2.0 (for most sizes) and can be run via the same tooling as Llama — Ollama, vLLM, llama.cpp. Alibaba also offers a hosted API through their DashScope platform. Qwen-VL variants support image input natively. Fine-tuning via Unsloth and Axolotl is fully supported.
Free to download and self-host. Hosted via Alibaba DashScope ($0.04-$0.80 per M tokens depending on size). Hosted via Together AI at standard Together pricing.
Most-deployed open model family in Asia. Used by Alibaba Cloud customers, Chinese enterprises, and developers worldwide who need strong multilingual support. Heavy use in the open-weight research community.