Qodo has launched Qodo Aware, a flagship product featuring the industry's first deep research agent for navigating enterprise-scale codebases, announced on September 10, 2025. Achieving an 80% accuracy rate on the DeepCodeBench benchmark, Qodo Aware outperforms leading AI coding agents like OpenAI Codex (74%), Claude Code (64%), and Gemini CLI (45%) in multi-repository tasks, addressing the critical challenge of context awareness in large codebases.
Qodo Aware introduces a deep research agent for enterprise codebases.
Achieves 80% accuracy on DeepCodeBench, surpassing Codex (74%), Claude (64%), and Gemini CLI (45%).
Features three modes: Deep Research, Ask, and Issue Finder agents.
Reduces investigation time from days to minutes across 100M+ lines of code.
Integrates with GitHub, GitLab, Bitbucket, and supports secure enterprise deployment.
Free MCP for individual developers to explore open-source repositories.
Modern software systems often span hundreds or thousands of repositories, making it challenging for developers to grasp dependencies and historical context. The 2025 State of AI Code Quality report notes that over 60% of developers find AI coding tools lacking in relevant context. Qodo Aware addresses this with a context engine that enables its deep research agent to navigate complex codebases in seconds, a task that could take engineers hours or days. “Developers don’t typically work in isolation, they need to understand how changes in one service affect systems across their entire organization and how those systems evolved to their current state,” said Itamar Friedman, co-founder and CEO of Qodo.
Qodo Aware offers three specialized modes:
Deep Research Agent: Conducts multi-step analysis across repositories for complex architectural queries and system-wide tasks.
Ask Agent: Delivers rapid responses for quick code-related questions via agentic context retrieval.
Issue Finder Agent: Identifies bugs, code duplication, and security risks across repositories.
These agents integrate with tools like Cursor and Claude Code, enhancing their context awareness, and are accessible via Qodo’s existing tools (Qodo Gen, Qodo Command, Qodo Merge) or as a standalone product through Model Context Protocol (MCP) and API.
On the new DeepCodeBench benchmark, which tests multi-repository coding tasks, Qodo Aware achieved an 80% accuracy rate, compared to 74% for OpenAI Codex, 64% for Claude Code, and 45% for Gemini CLI. Notably, Qodo Aware completed tasks in less than half the time of Codex, enabling faster development cycles. Qodo also released a multi-repository dataset to evaluate deep research agents, fostering further AI coding advancements.
Qodo Aware supports secure, single-tenant deployments for enterprises, ensuring code remains within organizational boundaries while integrating with GitHub, GitLab, and Bitbucket. For individual developers, a free Qodo Aware MCP enables deep research across popular open-source repositories. Enterprises can explore deployment options at qodo.ai/products/qodo-aware.
Qodo is an agentic code quality platform for reviewing, testing, and writing code, integrating AI across development workflows to strengthen code quality at every stage. The company’s AI agents are powered by a core platform that provides deep context awareness, enabling AI to understand the specific best practices of a codebase and solve complex coding challenges more effectively.
TechIntelPro empowers B2B technology decision-makers with the latest industry trends, deep insights, leadership perspectives and actionable intelligence to drive both organizational and professional milestones—and thrive on the tech pulse.