Coinbase and Binance are seeking access to Anthropic’s Mythos model as crypto exchanges and custodians prepare for new AI-driven cyber threats. Coinbase Chief Security Officer Philip Martin stated the model will accelerate both digital threats and defenses. Binance and Fireblocks are also assessing how Mythos could reshape cyber offense and defense tools.
This move follows Anthropic's introduction of Claude Mythos Preview, an AI model described as unusually capable at cybersecurity tasks. Anthropic claims the model can identify and exploit zero-day vulnerabilities in major operating systems and web browsers.
Mythos access is restricted to a select group of partners through Anthropic’s Project Glasswing. This initiative provides early access to companies like Amazon Web Services, Apple, Google, JPMorgan Chase, Microsoft, Nvidia, and Palo Alto Networks.
Earlier, Anthropic's Opus 4.6 model demonstrated its ability to find significant zero-day vulnerabilities in well-tested codebases. It reportedly uncovered over 500 high-severity flaws in open-source libraries, highlighting the urgency for crypto infrastructure firms to understand future AI capabilities like Mythos.