The Claude Fable 5 Export-Control Saga: A Full Timeline
From launch to suspension to restoration, the Claude Fable 5 story spanned less than a month. Here is the complete timeline with every key date.
TL;DR
Claude Fable 5's first month was extraordinary: it launched as the most capable generally available model, became the first commercial AI model pulled from the market by US export-control authority, and returned less than three weeks later. Here is the full timeline.
The Timeline
June 9, 2026 - Launch. Anthropic releases Claude Fable 5 and Claude Mythos 5. Fable 5 is the first publicly available Mythos-class model, state-of-the-art on nearly all tested benchmarks, priced at $10 / $50 per million tokens. June 12, 2026 - Suspension. At 5:21 PM ET, Anthropic receives a US export-control directive citing national security, following a jailbreak report from Amazon researchers. That evening, Anthropic disables Fable 5 and Mythos 5 for every customer worldwide - the first time export-control authority has been used to pull a commercial AI model. Mid-to-late June - Pushback. Researchers sign open letters, and public and political pressure mounts. Anthropic maintains the directive rests on a misunderstanding while complying with it. June 26, 2026 - Partial approval. The US government approves Claude Mythos 5 access for US organizations. June 30, 2026 - Controls lifted. The Department of Commerce lifts the export controls. Anthropic announces it will begin restoring access the next day. July 1, 2026 - Restoration. Fable 5 returns globally on the Claude Platform, Claude.ai, Claude Code, and Claude Cowork, with a new classifier blocking the reported jailbreak in over 99% of cases. Cloud re-enablement on AWS, Google Cloud, and Microsoft Foundry follows. July 7, 2026 - Usage credits. The transition window ends. Fable 5 moves off standard subscription allowances onto metered usage credits at $10 / $50 per million tokens, with Anthropic pledging to restore subscription inclusion when capacity allows.Why It Matters
In under a month, Fable 5 demonstrated both the promise and the fragility of frontier AI access. A single jailbreak report translated into a worldwide takedown, and restoration hinged on a targeted safety fix plus a regulatory reversal. For anyone building on frontier models, the episode is a lesson in concentration risk: capability alone does not guarantee availability.