Claude Fable 5 returns globally tomorrow
Anthropic is restoring the model following negotiations with the U.S. government. The rollout includes a new set of classifiers that block more cybersecurity-related tasks. Routine operations like coding and debugging will temporarily shift to Opus 4.8.
Anthropic is also working with Amazon, Microsoft, Google, and other Glasswing partners to build a unified framework for assessing AI jailbreak severity and response protocols. Other model providers are invited to join the effort.
The company is expanding its U.S. government collaboration on model testing, including pre-release access, data sharing on exploits and abuse, and joint research resources. The return of Fable 5 follows a lawsuit by Legion LegalTech against the Department of Commerce directive that disabled the model globally.
Up next
Claude Sonnet 5 increases performance at a higher cost per task
Claude Sonnet 5 hit 53 on the Intelligence Index, but task costs have jumped to $2.29. That is double the cost of Sonnet 4.6 and 15% more than Opus 4.8. Token prices are unchanged, but the model uses 40% more output tokens and triple the agentic steps per task.
Related stories
- Legion LegalTech sues US government over Fable 5 and Mythos 5 shutdown
- Nadella calls Claude Fable 5 too “editorially controlled”
- Four older Claude models get a watermark on November 16
- Claude text watermarks: detector limited to vetted groups
- Claude’s text now carries an invisible watermark
- No multiplayer in Claude Code Projects yet, Anthropic says
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.