Qwen
ACTIVEQwen (通义千问, Tongyi Qianwen) is a brand and AI model team developed by Alibaba Cloud (Alibaba Group). It releases open-weight large language and multimodal models (Qwen, Qwen2.5, Qwen3, Qwen-VL, QwQ) on Hugging Face and GitHub.
Qwen
Founded 2023 · Hangzhou, ChinaQwen (Chinese: 通义千问, Tongyi Qianwen) is a brand and team of artificial-intelligence models developed by Alibaba Cloud, the cloud unit of Alibaba Group. The first Qwen models were released in 2023 (beta as Tongyi Qianwen in April 2023, public release in September 2023). The team continuously publishes open-weight large language and multimodal models — successive generations include Qwen, Qwen2, Qwen2.5 and Qwen3, alongside specialised variants: Qwen-VL (vision-language), QwQ and QVQ (reasoning), Qwen2.5-Omni/Qwen3-Omni (omni-modal), Qwen-Image (image generation) and agentic tooling (Qwen-Agent, qwen-code). Many models are released under the Apache 2.0 license; others under the Tongyi Qianwen and Qwen Research licenses. Models are distributed on Hugging Face, GitHub and the ModelScope platform, and are also served through the Alibaba Cloud API. The project's principal technical lead was Junyang Lin (Tongyi Lab), who departed Alibaba in March 2026.
Founders
Chinese computer scientist and AI researcher (b. 1993), an early pioneer and the principal technical lead of the Qwen model family at Alibaba Cloud's Tongyi Lab.
Domain of activity
5 areas · 10 technologies · 2 industries- Language models
- Multimodal models
- Agent systems
- Content generation
- Computer Vision
- LLM
- Foundation Models
- Multimodal AI
- Generative AI
- NLP
- Agentic AI
- Speech AI
- Computer vision
- Tool Use
- Reinforcement Learning
- Software & Technology
- Research & science
Global presence
HQ Hangzhou · 2 regionsScale & funding
founded 2023Products & solutions
2 model families
Organizational relations
Profile & metadata
4 classification · 6 external linksClassification
- Statusactive
- Development stagemature
- Listed onNo
- Organization typeAI Lab
