Stop me if you’ve heard this one before: A little-known Chinese startup is making waves globally for an impressive new AI product.

No, we’re not talking about DeepSeek R1, the AI reasoning model that made waves among Western AI circles earlier this year. Instead, the hot new product du jour is called Manus, and it’s a new AI multipurpose agent — that is, more than an AI model, it’s an interface for controlling multiple models that can go off and, on its own, autonomously, complete complicated tasks like generating reports or running dozens of social media accounts on the user’s behalf.

If it sounds similar to the Deep Research modes offered by OpenAI, Google, and others, as well as OpenAI’s Operator agent and Anthropic’s Computer Use mode (the latter two of which can, like Manus, take control of a user’s computer or programs on it, moving the cursors and typing to perform actions within software), then congrats — you’ve understood what it aims to offer.

But for action-oriented leaders and decision makers within enterprises here in the West and abroad — such as chief technology officers, product managers, IT team leaders, and more — what do they need to know about Manus and the capabilities it offers? Read on to find out.

What is Manus and who’s behind it?

Manus AI was officially announced on March 5, 2025 on the social network X with a post from the company describing it as “the first general AI agent” that autonomously executes complex tasks rather than just generating ideas.

According to a South China Morning Post (SCMP) report, Manus was developed by a Chinese company called Butterfly Effect, which has offices in Beijing and Wuhan. The company reportedly has only a few dozen employees but has rapidly gained attention in China’s AI landscape.

The founding team includes entrepreneurs and experienced product managers, led by Xiao Hong, a 33-year-old serial entrepreneur and 2015 graduate of Wuhan’s Huazhong University of Science and Technology.

Xiao previously built WeChat-based applications that were acquired by larger companies and later launched Monica.ai, an AI assistant available as a browser extension and mobile app.

On its website, Manus is said to be named for the Latin word for “hand,” a nod to the fact that users can rely on it to perform actions for them, or, in my words, to “lend them a hand.”

How does Manus AI work?

Manus AI is designed as a multi-agent system, meaning it combines several AI models to handle tasks independently.

Unlike AI chatbots that assist users by providing information, Manus can research, analyze data, generate reports, automate workflows, and even write and deploy code.

According to X posts by Ji Yichao, co-founder and chief scientist of Manus AI, the system is built on Anthropic’s Claude 3.5 Sonnet — a nine-month-old AI model at this point — and fine-tuned versions of Alibaba’s Qwen models.

The team is currently testing upgrading Manus to Anthropic’s newest and most performant model, Claude 3.7, which is expected to further enhance its reasoning and execution capabilities.

Manus AI operates asynchronously, meaning users can assign tasks and walk away while it completes them autonomously. It is currently in private beta, with access granted through invitation codes.

How does Manus AI stack up against the U.S.-based competition?

One of the biggest reasons Manus AI has gained traction is its strong benchmark performance, beating U.S. firm OpenAI’s own o3-powered Deep Research agent and the “previous state-of-the-art” according to a graph posted on the official Manus website.

This claim, along with real-world tests, has led some AI power users and early adopters to the conclusion that Manus may be one of the most capable autonomous AI agents available today.

Beyond benchmarks, Manus has already proven itself on freelance platforms like Upwork and Fiverr and in Kaggle machine learning challenges, successfully executing complex real-world tasks.

AI influencers celebrate Manus’s arrival and impressive performance

Conversation about Manus in media and AI circles took off late last week when users on the social network X noted that some people were using it to automate the management of up to 50 social accounts at one time, in realtime, showing off its capabilities to create fleets of engagement that businesses could use for reviews.

In addition, though this hasn’t yet been proven for Manus, the same tech could presumably used for all kinds of marketing and influence campaigns, even political propaganda or disinformation.

But for the most part, AI power users and influencers in the West were largely impressed and celebrated Manus’s arrival — awed by their initial tests of it once they received some of the scarce beta invites, or observed the work of others with access to the tool.

Rowan Cheung, founder of The Rundown AI newsletter, described Manus AI’s launch as a potential turning point for AI agents and said “China’s second DeepSeek moment is here,” in a post on his LinkedIn account.

“This AI agent called ‘Manus’ is going crazy viral in China right now… It’s like Deep Research + Operator + Claude Computer combined, and it’s REALLY good.”

Cheung personally tested Manus and found that it:

• Created and deployed a biography website about himself, with 100 percent accuracy and real-time data retrieval.
• Found top rental spots in San Francisco based on crime rates, AI industry presence, and entrepreneurship density.
• Developed a full AI course, generating eight chapters of content, including tools, use cases, and prompts.

He received 500 invite codes from the Manus team and has been dolling them out to his subscribers and readers.

Former Googler and AI-focused YouTuber Bilawal Sidhu shared a hands-on video review, calling Manus “the closest thing I have seen to an autonomous AI agent.”

“It’s like you’re standing over the shoulder of somebody using a computer… asking them what to do at the highest level, and it basically does it for you.”

Sidhu tested Manus on various tasks, including:

• Researching locations—scanning Google Maps and news sources to recommend the best places based on regulations, accessibility, and safety.
• Developing video applications—automating video effects such as ASCII art and real-time filters.
• Extracting industry insights—pulling data from Reddit, Twitter, and other sources to generate comprehensive reports.

Sidhu emphasized that Manus breaks down complex workflows and executes them step-by-step, making it one of the most autonomous AI agents available today.

Another AI influencer, Chubby (@kimmonismus on X), claimed that Manus AI outperformed OpenAI’s Deep Research: “I do not get paid for promotion. I just tried preview access, and it outperformed OpenAI’s Deep Research by a lot! Mind-blowing.”

Andrew Wilkinson, co-founder of the startup exit provider and acquirer Tiny, wrote on X that “It’s absolutely insane. I feel like I just time travelled six months into the future. I threw it a zip file of 20 applicants for a CEO job and it did a deep dive on each, one by one…”

Big questions about Manus’s future remain

Despite the excitement, questions and concerns have emerged about Manus AI’s true capabilities.

According to the South China Morning Post, very few people have been able to test Manus AI because of severe server shortages following its sudden rise to fame.

Zhang Tao, Manus AI’s product partner, admitted in a social media post: “The current invite-only mechanism is due to genuinely limited server capacity at this stage… The team underestimated the enthusiasm of the public response, and our server resources were only planned for a demonstration level.”

This lack of availability has fueled speculation that Manus AI might be using scarcity marketing tactics to generate hype. Invitation codes for Manus were reportedly being sold on China’s second-hand marketplace Xianyu, leading to frustration among potential users.

In addition, users such as @teortaxesTex on X (the self-described “stan #1”, or obsessive fan, of DeepSeek) opined that “it’s a product devilishly optimized for influencers, which is why it exploded so much. Generating threadboy content, trip plans and such general interest [mind blown emoji, hand pointing down emoji] stuff – yah. STEM assistance, coding – worse than googling. More LLM than agent.”

In other words, some believe it’s good for influencers looking to build up a following through spammy lists and tips posts, but not as good as other tools at work like development. Even former VentureBeat writer and current TechCrunch-er Kyle Wiggers said in a piece over the weekend, “Manus probably isn’t China’s second ‘DeepSeek moment’.”

Is Manus AI the real deal?

Unlike DeepSeek R1, which developed its own foundation model, Manus AI relies on existing large language models (LLMs), such as Claude and Qwen. This has led some critics to question how much of Manus’s technology is truly original.

Yichao Ji acknowledged this in the official introduction video, stating: “Manus operates as a multi-agent system powered by several distinct models.”

Despite this, Ji emphasized that Manus AI is committed to open-source development and will release some of its models to the public later this year.

Currently in private beta, Manus AI is only available via invitation codes. The company plans to expand access soon, but server limitations may delay a full public launch. An open-source component is planned, allowing developers to integrate parts of its AI system.

While some view Manus AI as a breakthrough in autonomous AI agents, others remain skeptical about its true capabilities, business model, and scalability. As the hype around Manus AI continues to grow, the question remains: Is this the next DeepSeek moment, or just another overhyped AI product?