China's Manus takes on OpenAI with new Wide Research tool for massive multi-agent tasks

India Today

01-08-2025

There's an intense war going on in the tech world to build the fastest and most capable AI agents, outpacing the competition. From coding to image generation, tech giants are constantly launching new models to showcase the capabilities of their Large Language Models (LLMs). Right now, one of the most talked-about features is the Deep Research system. Industry heavyweights like OpenAI, Google, and xAI all are focusing towards AI agents that can methodically crawl through the web for hours to offer users detailed, in-depth reports. But taking a different approach, Chinese startup Monica has introduced the Wide Research tool for its Manus AI. And it is different. advertisementAccording to the company, rather than focusing on depth, Manus's Wide Research tool is designed to handle massive, high-volume tasks by deploying more than 100 AI agents in parallel. 'We affectionately call it Wide Research, but what it unlocks goes far beyond research. At its core, Wide Research is a system-level mechanism for parallel processing and a protocol for agent-to-agent collaboration,' the company wrote in an official blog post.How the Wide Research works At the heart of Manus's Wide Research tool is a special cloud-based virtual machine. This setup gives it, according to the company, 'Turing-completeness,' meaning it can perform nearly any kind of computing task. Whether it's data analysis, research, or design, users can simply chat with the AI agent to get things done.
The company explains that this new system understands natural language, so there's no need for coding or technical commands. Users can ask the agent to carry out complex tasks using plain instructions. What makes this setup powerful is its ability to scale computing resources as needed. According to the company, this AI tool can increase computing power up to 100 times more than the original offering. This makes it suitable for heavy tasks that would otherwise require high-end systems.In a short demo video, Manus co-founder and chief scientist Yichao 'Peak' Ji demonstrated how Wide Research can analyse 100 different sneakers at once. The tool simultaneously deployed 100 sub-agents, each evaluating a single shoe's design, pricing, and availability. The results were delivered as a sortable matrix in both spreadsheet and web formats in just minutes.But Wide Research isn't just for data collection. In another demo, Manus agents created poster designs in 50 distinct visual styles simultaneously, delivering fully finished design assets in a single downloadable ZIP file. The company explained that this flexibility stems from Wide Research's system-level parallel processing, which allows agents to communicate and coordinate in real time, without rigid role assignments.How Wide Research is different from Deep Research The capabilities of Wide Research contrasts with 'Deep Research' tools by companies like OpenAI, which typically use sequential, role-based approaches that focus on exhaustive, long-form investigations. Manus's Wide Research, on the other hand, prioritises scale and speed, making it especially useful for users who need to explore large datasets, compare hundreds of options, or generate a wide range of creative outputs at once.Manus Wide Research availabilityWide Research is rolling out today for users on the Pro plan (priced at $199 per month) and will gradually become available to those on the Plus plan ( priced at $39 per month) and Basic plan ($19 per month). Users on the Free plan will have access to a more limited version of the platform.- Ends