Project Glasswing: el modelo de IA de Anthropic que no puedes usar encontró 10.000 vulnerabilidades críticas en un mes
wwwhatsnew.com/2026/05/25/project-glasswing-claude-mythos-10000-vulnerabilidades-2026/
Anthropic to release Mythos-class models to the public
AI flaw-finder still under lock and key for now while company figures out guardrails, but extends access to more users including governments
www.theregister.com/security/2026/05/25/anthropic-to-release-mythos-class-models-to-the-public/5245596
Avec Claude Mythos, Anthropic détecte 10 000 failles – Le Monde Informatique
www.lemondeinformatique.fr/actualites/lire-avec-claude-mythos-anthropic-detecte-10-000-failles-100258.html