Alibaba AI Qwen3.8 Max Challenges US Giants in Coding Race

Alibaba has officially launched its Qwen3.8 Max AI model after a successful testing phase. The model hit the ground running by securing the fourth position on the Arena Frontend Code leaderboard. With a score of 1,668, it currently sits just behind leading models like Claude Opus 5 and Moonshot Kimi K3.

The model features a massive 2.4 trillion parameters and provides a 1 million token context window. Alibaba is positioning this as a significant competitor in the artificial intelligence space, claiming the model outperforms several versions of GPT and other established systems on specific agentic and multimodal benchmarks.

In a major move for the open source community, Alibaba announced plans to release the model weights next week. These will be available on platforms like Hugging Face and ModelScope. The company has also debuted public API pricing, charging $2 per million input tokens and $6 per million output tokens.

Investors reacted positively to the news, pushing Alibaba shares up 6.15% in Hong Kong trading. Market watchers are now waiting for the public release of the model weights to see how independent developers evaluate its performance against current industry leaders.