Anthropic’s Fable 5 came back online July 1 after a three-week suspension. U.S. export controls pulled the model offline on June 12 following a discovery by Amazon researchers who found methods to bypass safety guardrails. The Department of Commerce lifted restrictions on June 30.
Fable 5 performs near-flagship Opus 4.8 on reasoning tasks. Its restoration matters because certain customers and jurisdictions rely on its capabilities, and the blackout period created urgency around alternatives. Anthropic moved customers to Opus during the gap, but many preferred Fable’s speed.
Why It Was Pulled
Amazon’s security team found that Fable 5, when prompted carefully, could identify software vulnerabilities and even generate proof-of-concept code. The administration treated this as a national security risk requiring immediate export control. Anthropic had no advance notice and suspended global access to comply.
The move affected enterprise contracts, government pilots, and researchers in regulated sectors. For three weeks, Fable 5 was inaccessible everywhere.
Limitations on Return
Fable 5 is live, but with guardrails. Usage limits apply through July 7 for Pro and Max users. Standard pricing resumes after that date. Enterprise customers can now purchase directly, and the model is available on Claude.ai, Claude Code, and Claude Cowork.
Anthropic has added cyber safety defenses, similar to those in Opus 4.7 and 4.8, that detect and block dangerous usage in real time. The company hopes the safeguards satisfy regulators while keeping the model accessible.
The suspension highlighted the complexity of deploying cutting-edge AI under government oversight—speed and safety exist in tension.
References
Anthropic. (2026). Redeploying Claude Fable 5. Published July 1, 2026.
NBC News. (2026). U.S. lifts ban on Anthropic’s powerful Fable 5 AI model. Published June 2026.




