AI Agents Demonstrate Significant On-Chain Attack Capabilities

According to BlockBeats, recent research by Anthropic reveals that AI agents possess substantial on-chain attack capabilities. In simulated tests on smart contracts that were hacked between 2020 and 2025, Claude Opus 4.5, Sonnet 4.5, and GPT-5 collectively replicated vulnerabilities valued at approximately $4.6 million. Additionally, while scanning 2,849 contracts with no known vulnerabilities, the models discovered two new zero-day vulnerabilities and successfully simulated profits. The study highlights that the profitability of AI-driven on-chain attacks has doubled approximately every 1.3 months over the past year, indicating that AI technology is fully capable of autonomously exploiting vulnerabilities for profit.