HANGZHOU, China (Aug 2026) — Imagine leaving an artificial intelligence model to work on a software engineering project completely unsupervised for over two weeks. In internal testing, the new Alibaba Qwen3.8-Max AI did exactly that, autonomously executing a real-world software project over a 16-day period without any human intervention.
Alibaba recently rolled out Qwen3.8-Max, marking it as the most powerful model in its Qwen series to date. The multimodal foundation model brings massive upgrades to visual intelligence, research, coding, and long-horizon tasks. It boasts 2.4 trillion parameters and supports a context window of up to 1 million tokens.
Crushing the AI leaderboards
The new model is already making waves on global benchmarking platforms. It currently ranks fifth in the Text Arena and second in the Vision Arena. When it comes to autonomous coding, it holds the fourth spot in the Frontend Code Arena, proving its capability to execute long-term tasks entirely on its own.
Where can you try it?
Developers worldwide can now access Qwen3.8-Max via APIs on the Alibaba Cloud Model Studio. Alibaba plans to release the model weights next week, allowing for deeper customization. Everyday users can also experience its capabilities through QwenWork, the latest all-in-one workplace AI agent platform from Alibaba.
