BREAKING - JUNE 12, 2026
Claude Fable 5 and Mythos 5 are currently disabled for all users worldwide. The US government issued an export control directive at 5:21 PM ET on June 12 ordering Anthropic to suspend access for all foreign nationals. Anthropic disabled both models for every customer to ensure compliance. All other Claude models (Opus 4.8, Sonnet 4.6, Haiku 4.5) are unaffected. Anthropic disagrees with the directive and is working to restore access.
What Happened - The Full Timeline
Fable 5 launched on June 9, 2026 - less than four days before the directive arrived. It was the first public Mythos-class model, scoring 80.3% on SWE-Bench Pro and available to all Pro, Max, Team, and Enterprise subscribers at no extra cost until June 22. Stripe used it to migrate a 50-million-line codebase in one day. Anthropic had spent weeks in pre-launch testing with the US government, UK AI Safety Institute, and multiple third-party red teams.
At 5:21 PM ET on Friday June 12, 2026 — the same day SPCX began trading on Nasdaq — Anthropic received a letter from Commerce Secretary Howard Lutnick addressed to CEO Dario Amodei. The letter invoked national security authorities to issue an export control directive: suspend all access to Fable 5 and Mythos 5 for any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. The directive contained no specific details about the national security concern.
Anthropic's statement on anthropic.com/news/fable-mythos-access: "The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Anthropic models will not be affected. We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible."
What the Government Believes - A Narrow Jailbreak
The letter did not provide specific details of its national security concern. Anthropic's statement explains their understanding: "Our understanding is that the government believes it has become aware of a method of bypassing, or 'jailbreaking' Fable 5."
Anthropic explicitly disputes the severity of what the government found. From their full statement on anthropic.com:
"In the weeks leading up to the launch of Fable, Anthropic worked with the US government, the UK AISI, multiple private third-party organizations and internal teams to red-team Fable's safeguards for thousands of hours in total. These tests showed that Fable's safeguards are substantially more effective than those of any previously deployed model. No testers have yet been able to find a universal jailbreak — a jailbreak method that can very broadly bypass the model's safeguards, unblocking a wide range of cyber capabilities. We suspect that perfect jailbreak resistance is not currently possible for any model provider."
Anthropic's position is that any jailbreak the government has identified is narrow and non-universal — meaning it bypasses the classifier on a specific class of requests, not a broad capability bypass. The company further argues that a narrow jailbreak in Fable 5 is not meaningfully more dangerous than equivalent vulnerabilities in other models already available to foreign nationals, since Fable 5's safeguards are described as "substantially more effective" than any previously deployed model.
What Anthropic Disagrees With — and Why It Matters for the Industry
Axios reported that Anthropic "disagrees with the decision to retract the deployed commercial models due to the discovery of this narrow potential jailbreak." This is significant: Anthropic is publicly pushing back on a federal directive, framing it as a precedent concern rather than a compliance dispute.
The precedent argument matters. If the US government can retroactively shut down a deployed commercial AI model because a narrow jailbreak was discovered — especially one the company argues is less dangerous than capabilities already available elsewhere — it creates a new regulatory risk for every frontier AI lab. Any future model deployment could be reversed with a Commerce Department letter. Anthropic's pushback is not just about Fable 5: it is about whether this regulatory framework is appropriate for commercially deployed AI.
Axios described the move as "an escalation in Washington's effort to treat cutting-edge AI systems as national security assets" and noted that Anthropic "now finds itself on a Pentagon blacklist deeming it too dangerous for the government's own use, and in a Commerce Department licensing regime deeming it too dangerous for foreign use." That dual classification — too dangerous for the government AND for export — is unprecedented for a commercial AI company.
What This Means Right Now - Practical Impact
| Model |
Status as of June 13 |
API response |
| claude-fable-5 |
DISABLED — all users worldwide |
Error response |
| Mythos 5 (Glasswing access) |
DISABLED — all partners worldwide |
Error response |
| claude-opus-4-8 |
Available — unaffected |
Normal |
| claude-sonnet-4-6 |
Available — unaffected |
Normal |
| claude-haiku-4-5 |
Available — unaffected |
Normal |
| claude-sonnet-4-20250514 |
Available until June 15 only |
Normal — but deprecated Sunday |
For developers with Fable 5 in production: migrate immediately to claude-opus-4-8 as a fallback. Fable 5 is returning errors. This is separate from the June 15 deprecation of claude-sonnet-4-20250514 — both issues are active simultaneously. For the June 15 migration guide see our Claude API deprecation checklist ->
The Broader Context - Timing and IPO Implications
The timing is extraordinary. The directive arrived on the same afternoon that SpaceX began trading on Nasdaq — the first day of the SPCX IPO that was supposed to be the dominant AI capital markets story of the week. Anthropic itself is targeting an October 2026 IPO at a $965 billion valuation. A federal export control directive suspending its most advanced public model is the worst possible news for Anthropic's pre-IPO narrative.
The IPO implications are significant. Institutional investors evaluating Anthropic's October listing will now need to model regulatory risk in a way they did not before June 12. A Commerce Department that can shut down a deployed commercial model with a letter to the CEO — without specific details, citing a narrow jailbreak the company disputes — is a new category of business risk. Anthropic's S-1, when it goes public, will need to disclose this incident and the regulatory framework that enabled it.
For OpenAI, Google, and xAI: the same export control authority could in theory be applied to GPT-5.5, Gemini, or Grok. The Anthropic directive establishes a precedent that the Commerce Department can treat frontier AI models as export-controlled technology on short notice, without court review, citing national security. Every frontier AI lab's legal team is reading Lutnick's letter to Amodei right now.
Frequently Asked Questions
When will Fable 5 and Mythos 5 be restored?
Anthropic says it is "working to restore access as soon as possible" and "believes this is a misunderstanding." No timeline has been given. Restoration depends on resolving the dispute with the Commerce Department over the severity of the identified jailbreak. This could take days or weeks depending on whether Anthropic can demonstrate to the government's satisfaction that the vulnerability is narrow and non-universal.
Why was Fable 5 disabled for US citizens too if the order only covers foreign nationals?
Anthropic disabled it for all customers because there is currently no reliable technical mechanism to verify user nationality at the API level. Anthropic's statement: "The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance." Rather than risk inadvertently providing access to foreign nationals — including foreign national employees who might access the system through shared API keys or enterprise deployments — Anthropic took the conservative compliance position of disabling access universally.
What is an export control directive and how does the government have this authority?
Export control directives are issued under the Export Administration Regulations (EAR), administered by the Commerce Department's Bureau of Industry and Security (BIS). EAR controls the export of dual-use technologies — goods and technologies that have both commercial and military or national security applications. The government's position is that Fable 5 and Mythos 5 qualify as dual-use technology under EAR, similar to advanced semiconductor manufacturing equipment and certain encryption software. The authority to issue these controls does not require court approval and can be exercised on national security grounds without disclosing the specific intelligence basis.
What should I use instead of Fable 5 while it's disabled?
Claude Opus 4.8 (claude-opus-4-8) is the best available alternative. It scores 69.2% on SWE-Bench Pro — 11 points below Fable 5 but significantly better than GPT-5.5 (58.6%). All other Claude models are fully available. If you had Fable 5 hardcoded in your API calls, update to claude-opus-4-8 immediately — Fable 5 is returning errors and claude-sonnet-4-20250514 also retires this Sunday June 15.
Does this affect Anthropic's October 2026 IPO?
It is a significant negative development for Anthropic's pre-IPO narrative. The company was positioning Fable 5 as its flagship commercial product and the demonstration of its ability to safely deploy the most capable AI in the world. A federal suspension three days after launch — even a temporary one — creates regulatory risk that will need to be disclosed in the S-1 and addressed in the IPO roadshow. The October timeline is not necessarily at risk, but the story Anthropic tells investors will be more complicated than it was last week.
This story is developing. Check the Anthropic official statement for updates. For all AI news see the June 2026 daily AI news calendar.