Okay, here’s a rewritten version of the article, suitable for a blog post, keeping the same title:
**China’s AI Agent Manus Can Do Complex Tasks for You**
Meet Manus, a groundbreaking new AI agent from China that’s making waves globally. Developed by Butterfly Effect, a startup backed by Tencent, Manus is being hailed as the world’s first general AI agent, designed not just to chat, but to *do*. Unlike typical chatbots, Manus is built to handle complex tasks from start to finish.
As reported by *CNN Business*, Manus doesn’t just offer suggestions; it delivers complete results. In a private online demonstration, it showcased its capabilities by building websites, filtering resumes, and generating comprehensive property reports, all based on user-provided requirements.
**How Manus Works**
Peak Ji Yichao, co-founder and chief scientist at Butterfly Effect, explained in a post on X (formerly Twitter) that Manus leverages a combination of powerful large language models. This includes Anthropic’s Claude and fine-tuned versions of Alibaba’s open-source Qwen model. This hybrid approach allows Manus to tackle intricate reasoning and decision-making challenges that often stump other AI tools.
Transparency is a key feature. Manus includes a unique “Manus’s Computer” window. This allows users to see exactly what the AI is doing in real-time and provides the option to intervene if necessary. As *MIT Technology Review* put it, using Manus feels like collaborating with an exceptionally intelligent and efficient intern.
**Performance Testing**
*MIT Technology Review* put Manus through its paces, assigning it tasks like identifying tech journalists, searching for specific property types, and selecting candidates for their “Innovators Under 35” list. While Manus occasionally misinterpreted instructions or made hasty assumptions, reviewers found it easy to correct and noted rapid improvement with clear feedback.
A report by Chinese tech news site *36Kr* highlighted the cost-effectiveness of Manus. Each task reportedly costs around $2, significantly cheaper (nearly ten times) than similar tasks performed on OpenAI’s DeepResearch. This affordability could make Manus a game-changer for freelancers and small businesses – provided it can manage increasing user demand.
**Technical Challenges Remain**
Manus isn’t without its limitations. During *MIT Technology Review*’s testing, the system occasionally failed to load tasks due to high server traffic, displaying the message: *”Due to the current high service load, tasks cannot be created. Please try again in a few minutes.”* Accessing paywalled content also presents a challenge, and unclear instructions can lead to confusion.
To improve scalability and better serve its user base, Manus has formed a strategic partnership with Alibaba. The aim is to transition Manus to run entirely on Chinese AI models and computing infrastructure.