Manus AI, an artificial intelligence agent created by the AI company Monica, was unveiled last week. The firm, owned by the Chinese parent company ‘The Butterfly Effect,’ has labeled Manus as a general AI agent capable of managing various complex tasks. Following a video demonstration of the agent’s functionalities shared across social media, Manus earned the title of “China’s second DeepSeek moment.” This advanced system utilizes multiple large language models (LLMs) for its operations.
Introduction of Manus AI Agent
Detailed information regarding the Manus AI agent is accessible on its dedicated website. The company positions the agent as innovatively skilled in a diverse range of tasks applicable to both personal and professional settings. Unlike existing AI agents such as OpenAI’s Operator, Google’s Gemini Deep Research, and Salesforce’s Agentforce, which primarily focus on domain-specific functions, Manus aims to provide a more generalized approach.
According to Monica, Manus can undertake a variety of activities, including trip planning, stock analysis, research compilation, and online shopping. The agent operates within a sandboxed computer environment, where it performs tasks with cloud-based asynchronous processing. Yichao “Peak” Ji, a research lead on the Manus project, emphasized in the video that the agent draws on several AI models, although specific details about these models remain undisclosed.
Internal testing data suggests that Manus surpasses OpenAI’s DeepResearch tool in performance on the General AI Assistant (GAIA) benchmark. Presently, the AI agent is available only to selected testers, and there is no timeline for when it will be launched to the broader public.
Manus AI benchmark
Photo Credit: Manus
In the aforementioned video, the research lead mentioned that more information about the AI models would be forthcoming, aimed at benefiting the open-source community. Nonetheless, no specific details regarding the agent itself have been disclosed, raising questions about Monica’s claims regarding its capabilities.
Early testers of Manus have expressed mixed feedback about the AI tool. Victor Mustar, the Head of Product at Hugging Face, described it as “the most impressive AI tool I’ve ever tried” in a post on X (formerly Twitter). In contrast, Alexander Doria, the Co-Founder of the AI startup Pleias, reported experiencing frequent errors during his testing of the model.
