Anthropic Releases and Temporarily Suspends Claude Fable 5

Anthropic unveiled Claude Fable 5 on June 9, 2026, marking the first public release from its highly anticipated Mythos class of AI models. This new iteration was positioned as Anthropic's most capable model to date, specifically engineered for advanced, long-horizon agentic workloads. Fable 5 became accessible to a broad audience via the Claude API, as well as through major cloud platforms like AWS and Microsoft Foundry, indicating its readiness for widespread enterprise adoption. The model boasts impressive specifications, including a 1 million token context window and the ability to generate up to 128,000 output tokens per request. This significant increase in capacity and performance came with a corresponding price adjustment, with Fable 5 priced at $10 per million input tokens and $50 per million output tokens, effectively doubling the cost of its predecessor, Claude Opus 4.8. Anthropic emphasized Fable 5's robust safety measures, noting that extensive bug bounty programs, involving over 1,000 hours of testing, had failed to produce any universal jailbreaks, although the UK AI Security Institute had made some progress in early testing. Despite its promising launch and advanced capabilities, Claude Fable 5's public availability was short-lived. Just three days after its release, on June 12, 2026, Anthropic was compelled to temporarily suspend access to the model. This abrupt halt was a direct consequence of a United States government export directive. The suspension also extended to Claude Mythos 5, which shares the same underlying architecture as Fable 5 but was intended for a more restricted group of cyber defenders and infrastructure providers with certain safeguards lifted. The government's intervention reportedly stemmed from concerns flagged by Amazon's security team regarding a potential jailbreak in Fable 5, which was then escalated to the White House. White House AI adviser David Sacks indicated that the block might be temporary, expressing the administration's hope for Anthropic to remediate the safety issue and restore Fable 5 to general release as soon as possible. This incident underscores the escalating tension between rapid AI innovation and the critical need for robust safety and regulatory oversight, particularly for frontier models with significant capabilities. The event also highlighted a new risk factor for enterprises: model availability becoming subject to regulatory action without prior notice.