Just weeks after DeepSeek launched, China unveiled another potent AI model, Manus, an AI agent, highlighting the nation’s rapid entry into the AI race. Manus, developed by the Chinese startup Monica, is being compared to the leading AI systems developed by OpenAI, Google, and Anthropic. The company claims the model is a general-purpose AI that can perform tasks independently without human oversight.
Monica’s parent firm, ‘The Butterfly Effect’, is situated in China and has a generic artificial intelligence agent named Manus that can do a variety of complicated jobs. Manus received the title of “China’s second DeepSeek moment” when a video demonstrating the AI agent’s skills was circulated on social media. Interestingly, the system uses many large language models (LLMs) to function.
The Manus AI agent’s dedicated webpage has information about it. It “excels at various tasks in work and life,” according to the corporation, and is referred to as a “general AI agent.” Although there are a number of AI agents on the market, such as Salesforce’s Agentforce, Google’s Gemini Deep Research, and OpenAI’s Operator, they are mostly restricted to activities related to a given area.
Rather, the business asserts that Manus is capable of carrying out duties including research, stock and portfolio analysis, travel planning, and even internet shopping. It has a computer environment that is sandboxed so that it may execute tasks and analyze data asynchronously in the cloud. Yichao “Peak” Ji, a research lead on Manus, stated in the video that the AI agent is powered by many AI models, but she withheld any information about them.
The AI company also noted that Manus beats OpenAI’s DeepResearch tool on the General AI Assistant (GAIA) benchmark, according to internal testing. Only a select group of testers may now use the AI agent, and it is not yet clear when the platform will be open to the public.
In the video, the research lead also mentioned that the business will shortly provide further information about the AI models to the open-source community. Nevertheless, there is currently no information accessible or planned for public release regarding the AI agent itself. This has caused some people to question Monica’s claimed capabilities.
Since its March 6 debut, Manus has attracted international notice. Its developers claim that it performs better than OpenAI’s DeepResearch on the GAIA test, which gauges AI performance. Monica produced a demonstration video that shows Manus using the internet, collecting information, and carrying out intricate tasks in real time. It has the ability to create reports, spreadsheets, and presentations in addition to browsing websites and taking screenshots and recording online activities. Many have referred to this degree of automation as a significant advancement in AI technology.
Manus functions autonomously on the cloud, allowing it to carry out its designated activities even in the event that a user disconnects their device. Long-term initiatives can go without interruption because of these capabilities.
Manus actively explores the web, engages with websites, and shows its process in real time, in contrast to many AI models. This makes it easier for people to understand how the AI collects and analyses data.
It provides customized results by learning from user interactions. It adjusts to user preferences over time, making its replies more pertinent and high-quality.
To gather and interpret data, the AI may access platforms such as Telegram, X (previously Twitter), and others. As seen in its promotional video, it can even control several screens simultaneously. Manus produces more than simply text-based outputs. It can provide interactive presentations, comprehensive reports, and even code-based outputs like spreadsheets and data visualizations.
Several people who were given early access to Manus have expressed conflicting views regarding the artificial intelligence agent. In a post on X (previously Twitter), Hugging Face’s Head of Product, Victor Mustar, described it as the “most impressive AI tool I’ve ever tried.” But while testing the model, Alexander Doria, co-founder of Pleias, an AI business, said he frequently ran into problems.
With a lot more autonomy, Manus works similarly to AI chatbots like ChatGPT. Users just enter a task, like “Create a 7-day travel itinerary to Bali within a budget,” and Manus starts researching, gathering data, and structuring a response. The AI then compiles all pertinent information and presents a complete itinerary with links, maps, and travel suggestions. If the user disconnects, the AI keeps working in the cloud and notifies them when the task is finished.
Manus is now accessible via an online preview that is invitation-only. Monica mentioned that an official launch would occur soon, but she has not specified a date for the public release. In the upcoming months, the business also intends to make the concept open-source so that developers may include it into their own applications. This action may result in quick advancements and broad technological adoption.
Tech businesses are increasingly focusing on AI-powered agents and browsers, and Manus is leading the way in this trend. It signifies a move toward AI-driven productivity tools with its capacity to perform deep web research, provide structured outputs, and automate operations.
The enthusiasm around Manus indicates that AI innovation is picking up speed, especially in China. As AI systems get more proficient and independent, competition between firms like OpenAI, Google, and Monica will only get more fierce.
We use cookies to ensure that we give you the best experience on our website. If you continue to use this site we will assume that you are happy with it.
