- OpenAI and Paradigm released EVMbench, a benchmark for AI agents that find and fix Ethereum smart contract vulnerabilities.
- Top AI models can now exploit over 70% of critical bugs, a significant leap from less than 20% when Paradigm began the project.
- OpenAI expanded its Aardvark security agent and pledged $10M in API credits to bolster defensive crypto research.
- In market news, crypto majors fell, Coinbase’s Base layer-2 is ditching the OP Stack, and Brian Armstrong addressed quantum computing risks.
OpenAI and crypto research firm Paradigm jointly released a new AI-powered security benchmark on Tuesday, signaling a major integration between Artificial Intelligence and blockchain technology. Their EVMbench framework evaluates AI agents on detecting, patching, and even exploiting vulnerabilities in Ethereum smart contracts.
The benchmark draws on 120 real-world, high-severity vulnerabilities from audits, including scenarios from the security review of Stripe‘s Tempo blockchain. According to their joint release, model performance has improved dramatically, now exploiting over 70% of critical bugs compared to under 20% previously.
“It’s now clear to us that a growing portion of audits in the future will be done by agents,” Paradigm stated. OpenAI emphasized that measuring AI performance in economically relevant environments is critical as models become powerful tools.
This move directly tackles the over $5 billion drained from DeFi by exploits in recent years. However, the powerful capability also presents a dual-use risk, making advanced Hacking tools more accessible.
Meanwhile, cryptocurrencies traded lower amid geopolitical tensions. Coinbase CEO Brian Armstrong called the looming quantum computing threat a “solvable issue” for crypto. Consequently, the New York Stock Exchange president confirmed plans for a 24/7 blockchain trading platform.
In other news, Coinbase’s Base network announced it is dropping the OP Stack for its own unified client. Additionally, decentralized exchange MegaETH launched its World Markets platform to unify spot trading, lending, and perpetuals.
✅ Follow BITNEWSBOT on Telegram, Facebook, LinkedIn, X.com, and Google News for instant updates.
