Qodo Unveils High-Accuracy Deep Research Agent for Coding

Qodo has introduced Qodo Aware, a new deep research agent that achieves 80% accuracy on coding benchmarks, outperforming competitors such as OpenAI's Codex, which scored 74%, Anthropic's Claude Code at 64%, and Google's Gemini CLI at 45%, announced in a press release. This agent is designed to navigate complex, enterprise-scale codebases, providing developers with a comprehensive understanding of their organization's code.

Qodo Aware features a context engine that allows it to traverse multiple repositories, addressing the limitations of current AI coding tools that often miss relevant context. It includes three specialized agents: the Deep Research agent for complex architectural questions, the Ask agent for rapid code Q&As, and the Issue Finder agent for identifying bugs and security risks.

The platform uses advanced indexing and context retrieval techniques, enabling developers to modify complex systems without breaking dependencies. Qodo Aware is integrated into existing Qodo tools and is available as a standalone product, supporting platforms like GitHub, GitLab, and Bitbucket.