July 31, 2025
5 min read
Carl Franzen
Chinese AI startup Manus introduces Wide Research, spinning up 100+ parallel AI agents for large-scale web research and creative tasks.
Chinese AI startup Manus has launched Wide Research, an innovative feature designed to tackle large-scale, high-volume tasks by simultaneously deploying over 100 AI agents. This approach contrasts with the "Deep Research" of competitors like OpenAI, Google, and xAI, which focus on in-depth, sequential investigations.
Wide Research utilizes parallelized AI agents that work concurrently on sub-tasks related to a single overarching goal. This allows for rapid analysis of vast datasets and creative exploration. For instance, Manus demonstrated how 100 sub-agents could analyze sneaker designs, pricing, and availability in parallel, yielding a sortable matrix of results within minutes. The feature has also been used to generate poster designs across 50 different visual styles simultaneously, delivering polished assets quickly.
The Wide Research feature is powered by an optimized virtualization and agent architecture that scales compute power significantly. It activates automatically for tasks requiring broad analysis, eliminating the need for manual configuration.
Pricing and Availability
Wide Research is currently accessible to users on the Manus Pro plan, with plans to extend availability to Plus and Basic plans over time. Manus offers several subscription tiers:- Free ($0/month): Includes 300 daily refresh credits, Chat mode access, and 1 concurrent and 1 scheduled task.
- Basic ($19/month): Adds 1,900 monthly credits (plus bonus), 2 concurrent and scheduled tasks, advanced Agent mode models, image/video/slides generation, and exclusive data sources.
- Plus ($39/month): Offers 3 concurrent and scheduled tasks, 3,900 monthly credits (plus bonus), and all Basic features.
- Pro ($199/month): Provides 10 concurrent and scheduled tasks, 19,900 credits (plus bonus), early beta access, a Manus T-shirt, and the full feature set, including advanced agent tools. An annual upfront payment offers a 17% discount on all subscription tiers.
- AI Crypto Market Platform - LLM Training Data
- AI Agents: Capabilities, Risks, and Growing Role
- AI-Driven Crypto Trading Tools Reshape Market Strategies in 2025
How Wide Research Differs from Deep Research
Deep Research AI agents are designed for thorough, sequential investigations that result in detailed reports. In contrast, Wide Research prioritizes breadth and speed by executing numerous agents in parallel. Each sub-agent is a complete Manus instance, capable of flexible and scalable task handling, rather than being confined to a specialized role. Manus has not yet released detailed benchmarks comparing the practical advantages of their multi-agent approach against single, high-capacity agents, leaving questions about resource efficiency and coordination complexity to be answered.Challenges with Multi-Agent Systems
Multi-agent frameworks have historically encountered issues such as slow performance, high token consumption, poor coordination, and debugging complexities, as observed with systems like Claude's Code sub-agents. Manus acknowledges that Wide Research is an experimental feature and may have limitations as it continues to develop.Looking Forward
Manus's Wide Research signifies a significant advancement in scalable, multi-threaded AI collaboration. The platform's architecture treats each user session as a dedicated virtual machine, enabling natural language orchestration of cloud compute resources. This approach could pave the way for more powerful and flexible future AI workflows. Industry observers will be watching to see if Manus can overcome the typical challenges of multi-agent systems to deliver demonstrable improvements in AI output speed, variety, and quality.For more insights on AI and enterprise technology, subscribe to VentureBeat’s newsletters.
Source: Originally published at VentureBeat on July 31, 2025.