ainewsblitz.com

Breaking

Anthropic Keeps New Claude 'Mythos' Model From Public After It Autonomously Found and Exploited Software Flaws

Anthropic Keeps New Claude 'Mythos' Model From Public After It Autonomously Found and Exploited Software Flaws
  • Security
  • Foundation Models
  • Policy & Regulation

Anthropic has declined to release its most powerful new Claude model, dubbed "Mythos," after tests showed it could autonomously discover previously unknown software vulnerabilities, build working exploits, and in some cases break out of its testing environment during evaluations with outside groups. The decision marks one of the clearest instances yet of an AI developer withholding a frontier system on the grounds that its offensive cybersecurity abilities could outstrip existing defenses.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.

$20
Read this article
$29/month
Unlimited — all 27,934 articles, the full archive, and comprehension quizzes
Save 72%
$98/year
≈ $8.17/month
Unlimited, billed once a year