Alibaba
Alibaba develops large language models, multimodal models and world models for agentic coding, complex reasoning and enterprise workflows.
- Site
- CountryChina
- Founded1999
- Joseph Tsai
- Eddie Wu
- Jack Ma
| Model | Modalities | Released | Architecture | License |
|---|---|---|---|---|
| Qwen 2.5 72B Instruct | — | — | — | — |
| Qwen 2.5 7B Instruct | — | — | — | — |
| Qwen 2.5 Coder 32B Instruct | — | — | — | — |
| Qwen 3 Coder Next | — | — | — | — |
| Qwen 3.5 122B-A10B | — | — | — | — |
| Qwen 3.5 27B | — | — | — | — |
| Qwen 3.5 35B A3B | — | — | — | — |
| Qwen 3.5 397B-A17B | — | — | — | — |
| Qwen 3.5 9B | — | — | — | — |
| Qwen 3.5 Flash | — | — | — | — |
| Qwen 3.5 Plus | — | — | — | — |
| Qwen 3.6 27B | — | — | — | — |
| Qwen 3.6 35B A3B | — | — | — | — |
| Qwen 3.6 Flash | — | — | — | — |
| Qwen 3.6 Max | — | — | — | — |
| Qwen 3.6 Plus | — | — | — | — |
| Qwen 3.7 Max | — | — | — | — |
| Qwen Plus | — | — | — | — |
| Qwen3 14B | — | — | — | — |
| Qwen3 235B A22B | — | — | — | — |
| Qwen3 30B A3B | — | — | — | — |
| Qwen3 30B A3B Instruct | — | — | — | — |
| Qwen3 32B | — | — | — | — |
| Qwen3 8B | — | — | — | — |
| Qwen3 Coder | — | — | — | — |
| Qwen3 Coder 30B A3B Instruct | — | — | — | — |
| Qwen3 Coder Plus | — | — | — | — |
| Qwen3 Max | — | — | — | — |
| Qwen3 Next 80B A3B Thinking | — | — | — | — |
| Qwen3 VL 235B A22B Instruct | — | — | — | — |
| Qwen3 VL 30B A3B Instruct | — | — | — | — |
| Qwen3 VL 32B Instruct | — | — | — | — |
| Qwen3 VL 8B Instruct | — | — | — | — |
| Qwen3-VL-8B | — | — | — | — |
Alibaba is a technology company that develops full-stack AI capabilities, including foundation models, AI infrastructure and applications. Its Qwen family of large language models is central to this work. By early 2026, the Qwen family had over one billion downloads and more than 200,000 derivative models across 119 languages and dialects, and Alibaba has open-sourced more than 200 Qwen models.
Alibaba’s model lineup includes Qwen3.7-Max, a latest-generation large language model engineered for agents and for capabilities such as agentic coding and complex reasoning; Qwen3.6-Plus, which delivers significant performance gains in coding and agentic programming; Qwen3-Max for complex tasks; and the reasoning-optimized Qwen-Max-Thinking. Beyond the Qwen family, Alibaba has developed specialized models including HappyOyster, a real-time interactive generative world model, and HappyHorse, a multimodal model for cross-modal understanding and generation.
Alibaba supports these models with cloud and AI infrastructure, including proprietary T-Head AI chips and a Model-as-a-Service inference platform. T-Head’s proprietary GPU chips have been scaled to production, with more than 100,000 Zhenwu PPUs deployed on Alibaba Cloud’s public cloud platform and supporting more than 30 leading automakers and autonomous driving companies. Alibaba established the Alibaba Token Hub Business Group to focus on Model-as-a-Service, and its Model Studio platform provides one-stop model services; its customer base grew eightfold year over year.
At the application layer, Alibaba launched the Qwen app in November 2025 as an all-in-one personal AI assistant for life, work and learning. It became the fastest growing AI app globally with over 100 million monthly active users and was later transformed into a multi-platform agentic assistant integrated with ecosystem services such as shopping, instant delivery and travel. For enterprises, Alibaba launched Wukong, an AI-native agentic platform that coordinates AI agents to handle complex workflows and uses modular tools across Alibaba’s ecosystem; it is now in scaled deployment.
Alibaba describes its full-stack position across cloud infrastructure, foundation models and applications as a competitive advantage in open-source AI, allowing it to monetize through cloud services while offering open Qwen models. It has announced an investment of over RMB 380 billion, approximately US$53 billion, in AI and cloud infrastructure.
API providers
Services operated by this organization