Traders Bet 60% Odds Anthropic’s Claude Mythos Drops Today as Rumors Hit Fever Pitch

The model can autonomously discover zero-day vulnerabilities in real-world software, develop working proof-of-concept exploits, and complete complex, multi-step offensive and defensive cyber tasks that prior models could not. Partners using it under Project Glasswing, Anthropic’s defensive cybersecurity initiative, reported finding more than 10,000 high- or critical-severity vulnerabilities across critical software and open-source projects between April and May.

A June 8 Axios report added a sharper data point: Mythos Preview can turn a newly disclosed vulnerability into a working exploit in approximately 31 minutes, a process that previously took security researchers weeks.

Credible tech journalist Alex Heath and others reporting on X are pointing to an imminent wider rollout, possibly today or tomorrow. Key claims circulating include that any public version will carry substantial safety guardrails that significantly restrict offensive cybersecurity use cases, and that the model may launch under a consumer-facing name like “Claude Fable” or as the first model in a Claude 5-class tier.

On June 1 and 2, Anthropic expanded Project Glasswing access to ENISA, the EU cybersecurity agency, and approximately 150 additional organizations across more than 15 countries, including operators in power, water, and healthcare infrastructure. That expansion was still vetted access, not a public rollout.

Anthropic’s official blog and @claudeai and @AnthropicAI on X remain the primary confirmation points. The situation is fluid. Traders are clearly watching the same signals everyone else is, and the money is leaning hard toward this week.