Anthropic's latest research shows that AI Agents have considerable on-chain attack capabilities: in the simulation test of real hacked smart contracts from 2020 to 2025, Claude Opus 4.5, Sonnet 4.5 and GPT-5 together reproduced exploits worth about 4.60 million US dollars; while scanning 2849 contracts with no known vulnerabilities, the two models also dug out 2 new zero-day vulnerabilities and successfully simulated profits.
The study pointed out that AI's on-chain attack revenue has doubled approximately every 1.3 months in the past year, and the technology is fully capable of autonomous and profitable exploits.
AI Agents discover $4.60 million smart contract vulnerability in Anthropic simulation test
2025-12-02 03:05:56
Disclaimer:
1. The information provided does not constitute investment advice. Investors should make independent decisions and bear all risks themselves.
2. The copyright of this content belongs to the original author. The views expressed herein are solely those of the author and do not represent the stance or position of this website.
Previous article：
Anthropic模拟测试中AI Agents发现460万美元智能合约漏洞Next article：
HumidiFi：公募的FDV估值将为6900万美元
