Savant Chat, the top AI agent for Smart Contracts Auditing, has announced groundbreaking results that mark a paradigm shift in Web3 security. He won sixth place in the public Sherlock competition and competed directly with dozens of expert auditors. This groundbreaking event marks the first time AI has exposed performance comparable to professional human auditors in a fiercely competitive environment, reshaping the future of radically decentralized financial security.
Humans break through optimized barriers
The results are particularly noteworthy given that audit contests are designed essentially for human participants. According to professional auditors, the most challenging aspect of these competitions is not finding bugs, but proves that they are truly positive at the jury stage. Human auditors are fiercely competitive and rely heavily on explaining their findings in a way that withstands rigorous reviews. If inaccuracies or inadequate details are presented by either AI or human, then true positive findings are highly likely to be rejected during the review phase. Despite these important challenges, Savant Chat (ProPlan, an early glimpse into technology in June 2025) was excellent.
In addition to its importance, the symbiotic contract contests in which Savant Chat was tested were not characterized by critical or high-strength vulnerabilities. Moderately severe findings are usually identified by AI agents at times by the residual principle. However, AI’s ability to secure top positions in competitions focusing on less serious issues speaks volumes about its sophisticated analytical capabilities and adaptability.
Savant’s operational methodology includes deep analysis, generation of threat vectors, and formation of complex vulnerability hypotheses. A key final step – Creating proof of concept (POC) code to validate your findings and eliminate noise – The Savant Chat team strategically delegated this task to best-in-class open source SWE agents. This approach ensured the maximum signal from Savant’s core insights while efficiently verifying the findings.
Revolutionize the economics of security auditing
The economic impact is profound. In traditional audit contests, hypothesis generation consumes approximately 85% of the resources, but only 5% is required for verification. AI produces slightly more noise than the average human auditor, but is measured in multiples rather than orders of magnitude, but the cost of generating hypotheses with AI is orders of magnitude lower. “We see clear evidence that AI-driven hypothesis generation combined with human verification represents an effective and economically viable business model,” the team explained. “This is not about reducing costs, it’s about enabling more comprehensive security coverage across the Web3 ecosystem.”
Future: Cyborg Auditor, not an alternative
Although powerful, it is essential to acknowledge that AI is more restrictive and less effective than humans in identifying certain types of errors. However, SavantChat’s AI offers clear and invaluable benefits. Its perspective on bugs is very different from human auditors. This unique perspective allows AI to identify vulnerabilities that human experts may inadvertently overlook.
It has been revealed that 80% of zero-day exploits submitted to the Savant team were successfully identified and mitigated by Savantchat. This demonstrates that if these vulnerabilities were implemented through SavantChat, the majority of related hackings could have been prevented, highlighting their important and complementary capabilities. This positions AI not as an alternative, but as an essential enhancement to existing security measures.
The future of Web3 security is more than just arriving. It accelerates faster than ever.
About
Savant.Chat is an AI-powered Smart Contract security platform that automatically audits Solidity Code, developing vulnerabilities and integrating with CI/CD for real-time monitoring of developers, audit companies and businesses.
Tested in hundreds of real-world exploits. Savant Chat is a game-changing platform that instantly detects advanced threats such as broken encryption and economic flaws, allowing for fast, transparent and efficient audits. Savant Chat secures blockchain projects with AI that never sleeps.
