Chinese Firm Introduces Manus AI, a General Purpose AI Agent Said to Outperform OpenAI’s DeepResearch

As per its website, Manus AI can plan a trip, analyse stocks, convert lecture recordings into notes, and more.

Written by Akash Dutta, Edited by Siddharth Suvarna | Updated: 10 March 2025 13:07 IST

Highlights

Manus AI outperformed OpenAI’s DeepResearch on the GAIA benchmark
The AI agent is built using multiple AI models and systems
Manus can complete tasks asynchronously in the cloud

Chinese Firm Introduces Manus AI, a General Purpose AI Agent Said to Outperform OpenAI’s DeepResearch

Manus AI features a sandbox computer environment where it can complete tasks

Photo Credit: Manus

Manus AI, an artificial intelligence (AI) agent developed by the AI firm Monica, was introduced last week. China based 'The Butterfly Effect' is Monica's parent company, and called Manus a general AI agent that can handle a wide range of tasks of high complexity. After a video showcasing the capabilities of the AI agent was shared on social media platforms, Manus was given the accolade of “China's second DeepSeek moment.” Notably, the system operates using multiple large language models (LLMs).

Manus AI Agent Introduced

Information about the Manus AI agent is available on its dedicated website. The company calls it a “general AI agent” that “excels at various tasks in work and life.” While there are several AI agents available in the market, including the likes of OpenAI's Operator, Google's Gemini Deep Research, and Salesforce's Agentforce, these are largely limited to domain-specific tasks.

Instead, the company claims that Manus can complete tasks such as planning a trip, analysing stocks and portfolios, conducting research, and even purchasing products online. It comes equipped with a sandboxed computer environment where it can perform tasks and the processing occurs asynchronously in the cloud. In the video, Yichao “Peak” Ji, a research lead working on Manus highlighted that the AI agent is powered by several AI models, but did not disclose any details about them.

Based on internal testing, the AI firm also highlighted that Manus outperforms OpenAI's DeepResearch tool on the General AI Assistant (GAIA) benchmark. Currently, the AI agent is only available to some testers, and it is unclear when the platform will be accessible to the general population.

Manus AI benchmark
Photo Credit: Manus

The research lead also stated in the video that the company will soon share more details about the AI models to contribute to the open-source community. However, no details about the AI agent itself are available or have been announced to be made public. This has raised some concerns over the capability claims made by Monica.

Some of those who have received early access to Manus have shared contradictory opinions about the AI agent. Victor Mustar, the Head of Product at Hugging Face called it the “most impressive AI tool I've ever tried” in a post on X (formerly known as Twitter). However, Alexander Doria, the Co-Founder of the AI startup Pleias, claimed to have encountered errors on a regular basis while testing the model.

For the latest tech news and reviews, follow Gadgets 360 on X, Facebook, WhatsApp, Threads and Google News. For the latest videos on gadgets and tech, subscribe to our YouTube channel. If you want to know everything about top influencers, follow our in-house Who'sThat360 on Instagram and YouTube.