Alibaba Group has unveiled a new open-source artificial intelligence model, Qwen3-Coder, designed to assist with software development tasks.

Described by the company as its most advanced coding tool to date, the release marks Alibaba’s latest move in the intensifying global competition to develop cutting-edge AI technologies.

The model, launched on Wednesday, is optimized for tasks such as code generation, workflow management, and other complex programming processes. In a statement, Alibaba said Qwen3-Coder is particularly adept at “agentic AI coding tasks”—instances where the AI can work autonomously to solve programming problems with minimal human intervention.

Alibaba also shared benchmark results indicating that Qwen3-Coder outperforms several domestic rivals, including DeepSeek and Moonshot AI’s K2, in key coding capabilities. Additionally, the company claimed that its new model rivals top-tier international offerings, such as Anthropic’s Claude and OpenAI’s GPT-4, in specific performance categories.

The launch underscores the growing arms race among Chinese tech giants as they strive to match or surpass American firms in the development of generative AI. Companies on both sides of the Pacific are racing to release increasingly sophisticated models as AI becomes central to the next wave of global digital infrastructure.

Qwen3-Coder is being released as an open-source model, further positioning Alibaba as a key player in the broader push for accessible AI development tools.