Manus AI has introduced a groundbreaking research feature called Wide Research, allowing users to harness over 100 AI agents running simultaneously in parallel. This swarm-based architecture marks a shift away from the traditional deep research model used by OpenAI and others, instead opting for horizontal scalability that accelerates large-scale tasks. It’s a new take on how AI can multitask efficiently.
Each Manus agent operates within its own virtual machine, giving users access to massive computing power in a seamless conversational interface. Tasks like product comparisons, data analysis, and design generation are now handled by independent, collaborating subagents. With no fixed roles, each agent adapts flexibly to different sub-tasks, improving throughput for complex operations without requiring manual coordination or scripting.
Currently reserved for Pro-tier subscribers at $199 per month, Wide Research will gradually roll out to more affordable plans. Early users can request summaries, datasets, or visuals from multiple perspectives, and receive results packaged in files or tables. It’s a clear attempt to democratize high-performance computing, putting industrial-grade tools in the hands of regular users with simple prompts.
Still, there are concerns about reliability, hallucinations, and data privacy. Manus, which shifted operations to Singapore, has ties to Chinese tech firms that raise questions about jurisdiction and control. Critics also point out the lack of independent benchmarks proving that parallel swarms outperform deep-agent models in accuracy or insight. Despite this, the launch signals a bold step in AI design philosophy.
#AI #Manus #WideResearch #MultiAgentAI #ArtificialIntelligence #TechNews