The leak of the Anthropic model "Mythos" suggests a potential new era of cybersecurity competition and intensified attack and defense rivalry.

PANews reported on March 28th that, according to CoinDesk, Anthropic is internally testing its most powerful AI model to date, "Mythos." Notably, Mythos is described as posing "unprecedented cybersecurity risks," which has a direct impact on the crypto industry. As AI capabilities improve, its role in smart contract auditing, vulnerability discovery, and attack automation will be further amplified, potentially intensifying the DeFi offensive and defensive game. Recent events, including Ripple's AI security upgrade, the Ethereum Foundation's quantum-resistant security measures, and the Resolv stablecoin's loss of peg due to a contract vulnerability, all highlight the importance of security issues.

In the AI ​​race, this event may reshape industry benchmarks. While the decentralized AI network Bittensor recently released the Covenant-72B model and drove a surge in TAO's price, technological leaps by centralized labs like Anthropic may further widen the capability gap with decentralized solutions.

Anthropic stated that Mythos is currently only open for testing to "early adopters" and is not yet ready for a full release due to its high cost and sensitive nature. The company also acknowledged that the leak originated from a "human error" in its content management system and has removed the relevant public cache. This incident serves as a warning: even AI companies focused on cybersecurity can face significant data breach risks due to basic operational errors.