Qwen (通义千问) is a family of open-source large-language and multimodal AI models created by Alibaba Cloud.
Since its first public beta in April 2023, the project has grown from the original Tongyi Qianwen LLM into a broad ecosystem that now spans text, vision, audio and code-generation variants, all released under the permissive Apache 2.0 licence.
The latest Qwen 3 series (April 2025) scales up to 235 B effective parameters and supports a 128 K-token context window, multilingual reasoning across 119 languages and tool-calling for search, file analysis and voice chat. Qwen’s chat service at chat.qwen.ai integrates these capabilities in a single interface, enabling long-form conversation, document understanding, image and video analysis, code completion, and real-time voice dialogue.
Backed by Alibaba’s distributed training stack and open-source community on GitHub, Qwen aims to provide a competitive, transparent alternative to proprietary frontier models for researchers, developers and enterprises worldwide.
