Archive · 3 Aug 2026
Alibaba Unveils Qwen3.8‑Max, China’s Largest AI Model Today
Alibaba, a big Chinese tech company, made a new AI robot called Qwen3.8‑Max.
It can read and understand a lot of words, pictures, and videos all at once.
The robot is very big, with 2.4 trillion tiny parts that help it learn.
Alibaba says it can finish a big software job in just 16 days.
The robot will be available next week for people to use on Alibaba’s cloud platform.
This shows how China is trying to build the best AI tools.
Alibaba announced Qwen3.8‑Max, its largest AI model with 2.4 trillion parameters, close to Moonshot AI’s 2.8 trillion‑parameter Kimi K3.
The open‑weight model can handle text, images, and video up to 1 million tokens, using a mixture‑of‑experts design that runs only 95 billion parameters per request.
Qwen3.8‑Max topped Arena.AI’s Chinese text‑model leaderboard and ranked second globally for multimodal models, behind Claude Fable 5.
Alibaba claims the model finished a software‑engineering project in 16 days and will launch next week via Alibaba Cloud’s Model Studio.
The launch highlights China’s rapid AI race, with companies competing to build powerful yet cost‑efficient systems.
- Who
- Alibaba Group
- What
- Unveiled its Qwen3.8‑Max AI model
- When
- Monday, with release scheduled for next week
- Where
- China, announced on Arena.AI platform
- Why
- To compete in China’s fast‑moving AI race and offer a powerful, cost‑efficient open‑weight model
- Model name
- Qwen3.8‑Max
- Parameters
- 2.4 trillion
- Release date
- Next week (via Alibaba Cloud Model Studio)
- Company
- Alibaba Group
- Model type
- Open‑weight, multimodal (text, image, video)
- Token limit
- 1 million tokens