AI agents are getting good enough at finding attack vectors in smart contracts that they can already be weaponized by bad actors, according to new research published by the Anthropic Fellows program.
A study by the ML Alignment & Theory Scholars Program (MATS) and the Anthropic Fellows program tested frontier models against SCONE-bench, a dataset of 405 exploited contracts. GPT-5, Claude Opus 4.5 and Sonnet 4.5 collectively produced $4.6 million in simulated exploits on contracts hacked after their knowledge cutoffs, offering a lower bound on what this generation of AI could have stolen in the wild.
STORY CONTINUES BELOW
(Anthropic Labs & MATS)
The team found that frontier models did not just identify bugs. They were able to synthesize full exploit scripts,
Join the conversation!
Please share your thoughts about this article below. We value your opinions, and would love to see you add to the discussion!