The Open Source “Agentic Reasoning” Beats Google Gemini Deep Research

A new open-source framework called Agentic Reasoning, developed by University of Oxford researchers, is making waves in the field of deep research. This tool uses external LLM agents to tackle complex multi-step logical problems requiring extensive research. It outperforms leading RAG systems and closed-source LLMs on PhD-level scientific reasoning and domain-specific tasks. Notably, Agentic Reasoning surpasses Google Gemini Deep Research in open-ended Q&A tasks across finance, medicine, and law, demonstrating its potential to revolutionize AI-driven research capabilities.