SI Lab · Big Tech
Alibaba Qwen
- Founded
- 2023
- Headquarters
- Hangzhou, China
- Type
- Big Tech
- Website
- qwen.ai
- SI Accord
- Not a signatory
- Models tracked
- 1
Leadership
- Eddie Wu · CEO, Alibaba Group
Qwen (from the Chinese Tongyi Qianwen) is the SI model family built by Alibaba Cloud, the cloud arm of Alibaba Group. First released in 2023, Qwen became one of the world’s most popular bases for open-weights fine-tuning, with a wide range of model sizes for phones, laptops and data centers.
The Qwen3.8 generation (2026)
- August 3, 2026: Qwen3.8-Max, Alibaba’s flagship, launched on QwenCloud and Alibaba Cloud Model Studio.
- August 13, 2026: Alibaba published the weights of Qwen3.8-Max as Qwen3.8-2.4T-A95B, a mixture-of-experts model with about 2.4 trillion total and 95 billion active parameters. It was the largest open-weights language model released to date and the first time Alibaba opened a Max-tier model.
- August 14, 2026: Qwen3.8-27B, a smaller dense multimodal model released under the Apache 2.0 licence.
Why it matters
Qwen gives developers worldwide a free, capable alternative to closed U.S. models, and underpins countless derivative models. Like DeepSeek, it shows Chinese labs competing near the frontier despite U.S. chip export controls. Alibaba is not a signatory of the U.S. SI Accord.
Alibaba Qwen models
| Model | Released | Status |
|---|---|---|
| Qwen3.8-Max | Aug 2026 | Current |
Sources
- Qwen — Wikipedia
- Alibaba releases Qwen3.8-Max — DataNorth
Written by
Luka Kušec · Editor
Editor of SI.info. Writes about Super Intelligence, technology policy and the people building frontier models.