A team of software developers from China introduced a new AI agent called Manus AI, asserting its superiority over rivals like OpenAI. The announcement was made on social media last Wednesday.
Manus, developed by a Chinese company with the same name, reportedly outperformed OpenAI models across all three difficulty levels in the GAIA benchmark test. This test assesses general AI assistants on their capacity to solve real-world problems, as reported by Forbes magazine.
Manus operates independently across different domains, unlike traditional AI models that concentrate on chat-based workflows, according to Yichao ‘Peak’ Ji, co-founder and chief scientist of the program, in the video.
Various structures
Manus offers a valuable feature where it operates in the cloud asynchronously, enabling users to disconnect their devices while the AI continues working and delivers results later. It also possesses memory and evolves over time, adjusting to user preferences, as described by Yichao.
According to the Investng portal, Manus effectively chose and organized resumes, carried out detailed real estate investigations, and performed financial data analysis such as stock correlation studies.
DeepSeek impact
The launch of DeepSeek in late 2024, a Chinese AI model designed to rival OpenAI’s GPT-4, was considered a significant advancement in AI in China and generated increased activity among competing companies. In contrast, Manus, a distinct AI, functions as an independent agent capable of thinking, planning, and executing tasks autonomously, similar to an intern with unlimited focus.
Manus differs from Western models like ChatGPT-4 and Gemini by being able to function without human commands, initiating tasks independently, adapting to new information, and adjusting its approach as necessary. It is considered the first genuine general AI agent by Investng.
The program aims to anticipate and fulfill user needs, enabling professionals to concentrate on tasks involving creativity, strategy, and decision-making.
Manus AI’s effectiveness in collaboration is attributed to its capacity to learn, adapt, and cater to individual preferences, according to Investing.