What happened
On July 1, 2026, Anthropic restored Claude Fable 5 around the world. It came back across the Claude platform, Claude.ai, Claude Code and Cowork. The model had been suspended for 19 days. Anthropic added a new safety classifier. A classifier is a filter that spots and blocks certain requests. Anthropic says it stops the flagged jailbreak trick in over 99% of cases. A jailbreak is a trick that makes a model ignore its safety rules.
The context
The pause followed a report of a jailbreak method. The new filter is Anthropic's fix before bringing the model back.
What this means for you
👤EverydayThe most capable Claude is available again for everyday questions and tasks.
💼At workProfessionals regain a strong model for research, writing and coding.
🏢BusinessAccess is stable again, and the new filter lowers a known misuse risk.
🌍The worldThe quick pause-and-fix shows labs adding safety checks as they scale up.
Ledger entry: Release execution · Materiality 4 · T1 Primary → ▲ for Anthropic
Sources
- Anthropic, https://www.anthropic.com/news/redeploying-fable-5 T1 Primary