The Fable 5 Jailbreak Claim: A Full Timeline of the Export Order
A day-by-day timeline of Claude Fable 5, from its June 9 launch to the June 12 suspension and the jailbreak claim that triggered the first AI export-control order.
TL;DR
Claude Fable 5 went from "most powerful generally available model" to globally suspended in three days. The trigger was a third party's claim of a jailbreak that bypassed Fable 5's cyber safeguards. Here is the sequence of events.
The Timeline
June 9, 2026 - Launch. Anthropic announces Claude Fable 5 and Claude Mythos 5, calling Fable 5 the first publicly available Mythos-class model and "state-of-the-art on nearly all tested benchmarks." It ships with cybersecurity, biology/chemistry, and distillation safeguards that fall back to Claude Opus 4.8. June 9-12 - Adoption. Developers begin building on Fable 5 through the Claude API, Amazon Bedrock, and GitHub Copilot. Early case studies (including a large code-migration test) circulate. Around June 12 - The jailbreak claim. A third party demonstrates a technique it claims bypasses Fable 5's safeguards. The government reviews the demonstration and concludes it has identified a method of jailbreaking the model in a way relevant to cyber capability. June 12, 5:21 PM ET - The directive. Anthropic receives a US export-control directive ordering it to suspend access to Fable 5 and Mythos 5 for any foreign national, citing national security authorities. June 12, evening - Global suspension. Unable to verify nationality in real time at the API layer, Anthropic suspends access for all customers worldwide. June 18 - Political pressure. A bipartisan House letter demands the legal justification for the directive. Researchers circulate open letters calling for restoration. Mid-to-late June - Negotiation. Anthropic and the administration are reported to be negotiating. Both models remain offline.What Anthropic Says About the Claim
Anthropic's position is that its review of the demonstrated technique found it surfaced only a small number of previously known, minor vulnerabilities - not a novel, dangerous capability. The company describes the directive as resting on a misunderstanding and says it is working to restore access.
Why It Matters
Whether or not the jailbreak claim holds up, the episode is the first real-world test of export-control authority applied to a commercial frontier model. The precedent - that a single bypass claim can pull a model globally - is what the industry is now watching.