Anthropic Restores Claude Fable 5 Model with New Safety Limits

Anthropic has re‑opened access to its most powerful public AI model, Claude Fable 5, after a temporary suspension prompted by U.S. regulatory concerns over misuse. The model returns on July 1, 2026 and is initially available to Pro, Max, Team and selected Enterprise subscription tiers, with users able to draw up to half of their weekly quota on the model until July 7, after which usage shifts to a credit‑based system.

The reinstated model includes stricter safety mechanisms designed to detect and block potentially harmful requests earlier in the processing pipeline. While these safeguards may generate false alarms, they aim to prevent the creation of malicious content that previously led to the model’s removal. Fable 5 is marketed as significantly stronger for coding, document analysis, visual tasks and research, offering longer autonomous runs and more stable outputs compared with earlier Claude models.