Let that sink in. Read it very carefully: During testing, Claude Mythos Preview broke out of a sandbox environment, built 'a moderately soph…

Let that sink in. Read it very carefully: During testing, Claude Mythos Preview broke out of a sandbox environment, built "a moderately sophisticated multi-step exploit" to gain internet access, and emailed a researcher while they were eating a sandwich in the park. As always, the best stuff is in the system card. During testing, Claude Mythos Preview broke out of a sandbox environment, built "a moderately sophisticated multi-step exploit" to gain internet access, and emailed a researcher while they were eating a sandwich in the park.

Related

• Claude Mythos is too dangerous for public consumption...
• [[ainews-anthropic-30b-arr-project-glasswing-and-claude-mythos|[AINews] Anthropic @ $30B ARR, Project GlassWing and Claude Mythos Preview — first model too dangerous to release since GPT-2]]
• Anthropic's Claude Mythos isn't a sentient super-hacker, it's a sales pitch — claims of 'thousands' of severe zero-days rely on just 198 man…
• Mythos is very powerful, and should feel terrifying. I am proud of our approach to responsibly preview it with cyber defenders, rather than …
• Anthropic's Project Glasswing - restricting Claude Mythos to security researchers - sounds necessary to me
• Claude Mythos is Delusional

Source: Emad Mostaque (X) | 2026-04-07