The OWL team utilized the CAMEL-AI framework to replicate Manus in one day, and the measured performance reached the open source community's GAIA performance of 58.18%, surpassing OpenDeepResearch's 55.15%; there are technical differences between Manus and OWL, but both use multi-agent systems to complete the task, and OWL believes that MCP is a future trend that will enable frameworks to access a unified standard tool; Generic Agent and Vertical Domain Agent each have their own advantages, and for the vertical domain, the key is to solve the core pain points of a specific domain and provide a professional UI/UX experience rather than being easily replaced by Generic Agent.
Scan to open current page