China has introduced Manus, a state-of-the-art AI agent, claiming it surpasses OpenAI’s models in performance, according to Shanghai Securities News.

Manus achieved a top score in the GAIA benchmark test, positioning it as a highly advanced general-purpose AI assistant.

Unlike traditional AI models, Manus is designed for both idea generation and real-world execution, allowing it to complete complex tasks autonomously. A demonstration highlighted its ability to write reports and create tables with tailored outputs. The AI agent also features a highly adaptive “digital brain”, enabling independent learning, complex instruction processing, and cross-domain collaboration.