Forget DeepSeek—Manus AI is China’s boldest challenge to OpenAI yet

A new player in the AI landscape, Manus, is making waves with its promise of full autonomy in task execution. Developed by Chinese startup Monica, this AI agent is designed to go beyond text generation, delivering complete, actionable results. Since its launch on March 6, Manus has been compared to leading AI systems from OpenAI, Google, and Anthropic, with its creators claiming it outperforms OpenAI’s DeepResearch on the GAIA benchmark.

What sets Manus apart?

Unlike traditional AI models that provide responses based on prompts, Manus functions as an independent agent that can plan, execute, and finalize tasks with minimal user intervention. It is designed to handle real-world complexities autonomously, making it a step ahead of conventional AI chatbots.

For instance, when asked to generate a research paper on global warming, Manus does not just draft content. It researches credible sources, creates interactive charts, codes visual elements, and compiles everything into a structured format—eliminating the need for multiple user inputs. The same applies to stock analysis, course creation, and travel planning.

A demonstration video from Monica showcases Manus in action, seamlessly collecting data and performing tasks in real time. Early testers claim it surpasses existing AI research tools in automation and execution.

Key features

1. Fully Autonomous Task Execution

• Manus continues working in the cloud, even if the user logs off.

• Demonstrations have shown the AI navigating through multiple screens, gathering information from platforms like X (formerly Twitter) and Telegram, and taking screenshots to support its findings.

2. Real-time Interaction & Transparency

• Users can observe Manus in action as it browses the web, accesses tools, and compiles results step by step.

• In one example, Manus created a complete travel itinerary for Japan, displaying each phase of research in real time.

3. Personalized Adaptation

• The AI learns from user interactions, refining its output to align with individual needs and preferences over time.

4. Background Processing

• Once assigned a task, Manus operates independently and sends a notification upon completion, even if the user is offline.

5. Multi-Format Output

• Beyond text generation, Manus interacts with web pages, tracks activity, and produces structured outputs like PDFs, spreadsheets, and presentations.

Availability & access

Currently, Manus is available on an invitation-only basis, with no official public release date announced. However, growing interest suggests that a broader launch may be forthcoming. Monica has also stated plans to open-source the model, allowing developers to integrate Manus into external applications.

Using Manus is similar to engaging with AI platforms like ChatGPT or Grok. Users input a request—such as crafting a budget-friendly travel plan for Bali—and Manus autonomously researches, organizes data, and delivers real-time, structured results.

With its advanced autonomy, Manus is positioning itself as a major contender in AI evolution. While competition remains fierce among OpenAI, Google, and Anthropic, Manus’ ability to independently execute complex tasks could mark a turning point in AI development. Whether it will redefine the AI landscape remains to be seen, but it has undoubtedly set a new precedent for autonomous AI systems.