Anthropic introduces Covered Models to mitigate risks
The company now designates models as Covered Models once they hit capability thresholds in engineering, agentic workflows, scientific research, or cybersecurity. This status triggers enhanced security measures and special data handling rules.
For Covered Models, Anthropic applies policies that analyze users' query histories, not just individual prompts. This helps detect abuse that only becomes apparent from patterns across multiple actions.
The list already includes two models effective June 9, 2026. Claude Fable 5 is available to all users via Claude, Amazon Bedrock, Google Cloud, and Microsoft Foundry. Claude Mythos 5 has limited access and is only provided to approved partners.
Related stories
- Insiders say Anthropic oversold the rogue AI scare
- Federal Register searched comments with Alibaba's Qwen
- Anthropic's Jack Clark wants a kill switch others can check
- Amodei wants a speed limit on AI self-improvement
- UK's AI testing agency didn't get Anthropic's latest model
- Anthropic paused high-risk AI testing
Comments
No comments yet. Be the first.
Join the conversation
Sign in with Google to leave a comment. Your name and avatar come from your Google profile, and the comment appears after moderation.