TMTPOST – A Chinese research team has officially launched "Manus," a general AI Agent designed to go beyond idea generation and take action to solve real-world problems. Manus has achieved state-of-the-art (SOTA) performance in the GAIA benchmark, surpassing OpenAI's models in the same category, according to the team. Unlike traditional large language models, Manus is positioned as a highly capable assistant that not only provides insights but also executes tasks, making it a more practical and action-oriented AI solution.