Claude Restores Fable 5 Access After Government Cybersecurity Review.

Anthropic has restored access to its Fable 5 model after updating cybersecurity safeguards in consultation with the US government. The company announced the re-deployment on July 1, weeks after initially launching the model with selective safety guardrails.

The return follows what Anthropic described as conversations with government officials about strengthening protections for sensitive capabilities. Users can now access Fable 5 again, though the updated safeguards will trigger slightly more false positives than the previous version in the near term.

When the updated classifiers flag a request, users receive a clear notification and get a response from Claude Opus 4.8 instead. The vast majority of coding work remains unaffected, according to Anthropic.

Access limits for paid plans:

• Available through July 7, 2026
• Users can allocate up to 50% of weekly usage limits to Fable 5
• After hitting the 50% threshold, users must switch to another model
• Continued access via usage credits remains an option

The biology and chemistry classifiers remain unchanged from the initial Fable launch. These safeguards still cast a wider net than ideal, triggering fallbacks to Opus 4.8 on basic biology-adjacent questions. Anthropic said improvements to these classifiers will land soon.

Anthropic is asking users to report mistaken flags to help refine the classifiers. In Claude Code, users can run /feedback to file a report. On Claude.ai and Cowork, feedback flows through the thumbs buttons.

This feedback loop aims to reduce false positives over time as the company tunes the safeguards. The approach mirrors Anthropic's broader strategy of iterative deployment, where models launch with conservative restrictions that loosen as the company gathers real-world data.

The timeline is tight. With access guaranteed only through early July for paid plans, Anthropic will need to refine the classifiers quickly if it wants to maintain user confidence in Fable 5's utility for everyday work.

The redeployment comes amid growing scrutiny of frontier AI models by US regulators. Anthropic thanked partners across government, industry, and research for working alongside the company to restore access.

The episode highlights the balancing act facing AI labs as they deploy increasingly capable models. Anthropic pulled Fable 5 shortly after launch to address government concerns, then spent weeks updating safeguards before bringing it back online.

For users who depend on Fable 5 for coding and other tasks, the interruption underscores the regulatory uncertainty still shaping access to cutting-edge AI. The model's return with tighter guardrails suggests this trade-off between capability and control will continue as models grow more powerful.