Entering China
In the early morning of March 6th, it was another sleepless night in the tech circle after DeepSeek, and everyone was swiped by a product called Manus.
At the same time as Apple released the new product, many people stayed up all night to use the invitation code for Manus, the world’s first AI agent product developed by Monica.im.
According to his team, Manus is a truly autonomous AI agent capable of solving a wide range of complex and changing tasks. Unlike traditional AI assistants, Manus can not only provide suggestions or answers, but also deliver complete tasks directly.
Manus comes from the Latin Mens et Manus, which means mind and hand.
Manus uses a multisig system that is driven by multiple independent models. Later this year, there are plans to open-source some of these models, especially the postering part of Manus.
Simultaneously with Manus, there is also a four-minute demo. In these cases, Manus completes the entire process from planning to execution completely autonomously, demonstrating the true agent capability rather than a simple assistant function.
For example, in the first task of selecting candidates from 15 resumes for the position of reinforcement learning algorithm engineer, Manus has shown himself to be like a human intern, manually unzipping files and going through each resume page by page, while recording important information in it. Officials said that the content shown was only the tip of the iceberg of Manus’s capabilities.
When the AI Agent finally outputs a complete and professional result by invoking a long chain of thought and tool calls, users begin to sigh that AI can really help humans do things.
To ensure the reproducibility of the results, Manus is evaluated using a configuration that is exactly the same as its official version.
According to the official website, in the GAIA benchmark (which evaluates the ability of a general-purpose AI assistant to solve real-world problems), Manus achieved a new state-of-the-art (SOTA) performance across all three difficulty levels.
Xiao Hong, the founder behind Manus AI, graduated from Huazhong University of Science and Technology with a major in software engineering.
After graduation, he continued to start his own business, founded Nightingale Technology in 2015, launched Yiban Assistant and Weiban Assistant, served more than 2 million B-end users, and received investment from Tencent and Zhen Fund.
He also developed Monica, an AI assistant product called All-in-One, which was initially launched as a browser add-on. By integrating with mainstream large models (such as Claude 3.5, DeepSeek, etc.), Monica provides functions such as chat, translation, and copywriting. In the early days, Monica mainly focused on overseas markets, with more than one million users, and became the leading product in the field of AI plug-ins. In February this year, the Chinese version of Monica has been opened for internal testing and is currently free for domestic users. This version is based on the DeepSeek R1 and V3 models, which has deep reasoning and thinking capabilities, and supports memory functions and real-time online search.
Manus’ philosophy of technology, “less structure, more intelligence”, is also a bit different from the mainstream. They believe that when the data is good enough, the model is strong enough, the architecture is flexible enough, and the engineering is solid enough, capabilities such as computer use, deep research, and coding agent will naturally emerge without the need to be designed as a specific product feature.
On March 6, OpenAI announced that it would charge $20,000 (about 145,000 yuan) per month for AI Agents who have reached the doctoral level, mainly for the high-end needs of enterprise users, especially in data-intensive industries such as finance, medical care, and manufacturing.
