\n<\/p>\n
Claude Fable 5 app logo is seen on a smartphone against a Mythos Class background in this illustration photo taken in Italy on June 10, 2026. Anthropic launched Claude Fable 5 on June 9, 2026, as the first publicly available model in its Mythos-class tier, which exceeds the capability of the Opus line. The full Mythos 5 model remains accessible exclusively to partners participating in Project Glasswing. (Photo Illustration by Matteo Della Torre\/NurPhoto via Getty Images)<\/span><\/p>\n NurPhoto via Getty Images<\/small><\/div>\n<\/div>\n<\/figure>\n On the evening of 12 June 2026, three days after Anthropic had launched its most capable AI models to the public, the U.S. government ordered them taken down.<\/p>\n According to Axios<\/u><\/a>, Commerce Secretary Howard Lutnick sent a letter to Anthropic CEO Dario Amodei directing the company to suspend all access to Claude Fable 5 and Claude Mythos 5 by any foreign national, anywhere in the world, including foreign nationals working inside the United States. Soon enough, both models were offline for every customer on the platform worldwide.<\/p>\n X (Formerly Twitter)<\/span>Anthropic on X (formerly Twitter): “The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.The net effect of… \/ X”<\/span><\/span><\/a><\/p>\n The move appears to be among the most aggressive uses of export-control powers against a commercially deployed AI model.<\/p>\n If you\u2019ve been using Claude in the last week, you may have noticed two new model names. Fable 5 was the headline launch \u2014 Anthropic\u2019s first general-release version<\/u><\/a> of its Mythos model family, which the company had previously described<\/u><\/a> as too powerful in the cybersecurity domain to release publicly. Mythos 5, the version with fewer restrictions, was available only to a vetted group of organizations through Anthropic’s Project Glasswing program.<\/p>\n Both are now gone. <\/p>\n \u201cThe US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees,\u201d Anthropic said in its official statement<\/u><\/a> posted shortly after the order arrived Friday evening. \u201cThe net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.\u201d<\/p>\n All other Claude models \u2014 Opus, Sonnet, Haiku \u2014 remain available.<\/p>\n<\/section>\n The trigger seemed to be a reported technique for bypassing Fable 5’s safety guardrails \u2014 the filters used to stop the model from accessing the advanced cybersecurity capabilities baked into the underlying Mythos architecture. The concern was that if those guardrails can be defeated, a consumer-facing AI product effectively becomes an unrestricted cyber tool.<\/p>\n According to Axios<\/u><\/a>, the administration had tried to stop Anthropic from releasing the models at all, and failed. The export control order was the government’s next move.<\/p>\n A day after the shutdown, White House AI adviser David Sacks laid out the administration\u2019s account in a post on X. <\/p>\n \u201cA highly credible trusted partner of both Anthropic and the USG who was testing Fable came forward with a jailbreak of those guardrails,\u201d Sacks wrote. \u201cThe Admin asked Dario to fix the jailbreak or de-deploy the model. Dario refused.\u201d He called the administration\u2019s response \u201creluctant\u201d and said resolution was straightforward. \u201cThe ball is in Anthropic’s court.\u201d<\/p>\n X (Formerly Twitter)<\/span>David Sacks on X (formerly Twitter): “I’ve had a number of conversations with folks inside and outside government about the current situation with Anthropic, and here is what I believe to be true: – As we know, Anthropic publicly released its Mythos class models earlier this week under the commercial name Fable…. \/ X”<\/span><\/span><\/a><\/p>\n Anthropic broadly disputes these accounts \u2014 both the severity of the jailbreak and the characterization of its response. In its public statement, the company said it had reviewed what it believes to be the underlying report and concluded, \u201cthe level of capability demonstrated is available from other publicly deployed models, including OpenAI\u2019s GPT-5.5, and is used by cybersecurity defenders routinely.\u201d<\/p>\n \u201cWe disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people,\u201d Anthropic said<\/u><\/a>. \u201cIf this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers.\u201d<\/p>\n A source close to the company reportedly<\/u><\/a> contended that Anthropic was not presented with any details and never refused to fix issues. The two accounts have not been reconciled as of this writing.<\/p>\n<\/section>\n The identity of the \u201ctrusted partner\u201d who flagged the jailbreak has not been officially confirmed, but reporting by Semafor<\/u><\/a> and the Wall Street Journal<\/u><\/a> points to Amazon. <\/p>\n CEO Andy Jassy reportedly told Treasury Secretary Scott Bessent and other government officials<\/u><\/a> that Amazon researchers used Anthropic\u2019s Claude Fable 5 to obtain information that could be used in cyberattacks. The government subsequently imposed the export control ban.<\/p>\n Amazon is one of Anthropic\u2019s largest investors<\/u> and provides much of its cloud infrastructure, making its purported role in triggering the shutdown one of the more striking and interesting elements of the story. <\/p>\n \u201cIt\u2019s not uncommon for governments to seek our counsel on potential security risks,\u201d an Amazon spokesperson told Semafor, declining to share specifics.<\/p>\n Another angle reported<\/u><\/a> is that the White House acted partly over suspicion that a China-linked group had accessed Mythos, raising fears that the model could be reverse-engineered or distilled by a foreign adversary. <\/p>\n<\/section>\n Since early 2025, Anthropic and the Trump administration have been locked in a thorny conflict<\/u>. <\/p>\n The Department of Defense previously designated Anthropic a supply chain risk following the collapse of talks between the two, and Anthropic filed a lawsuit challenging that designation \u2014 a case still active in court. <\/p>\n \u201cThree months ago, @DeptofWar kicked @AnthropicAI out of our building \u2014 forever,\u201d Defense Secretary Pete Hegseth posted on X earlier this week. \u201cEvery passing day proves why that was the right move.\u201d<\/p>\n Meanwhile, Sacks pushed back on the idea that old grievances were driving the export control order. \u201cThose trying to misdirect and tie this action to the prior DoW\/Anthropic issues are wrong,\u201d he wrote on X.<\/p>\n There\u2019s also a notable timing detail. Just one day after launching Fable 5, Amodei published a major policy essay calling on the U.S. government to hold legal authority to block or reverse frontier AI models that fail independent safety testing. Two days later, the government used that authority against Anthropic.<\/p>\n The shutdown also lands at a commercially sensitive moment. Anthropic filed a confidential IPO prospectus with the SEC earlier this month, disclosing<\/u><\/a> a revenue run rate of $47 billion and a valuation of $965 billion.<\/p>\n Now that a sitting government has shown it can switch off a widely used AI product, mid-deployment, on the basis of a security assessment that the company says is inaccurate, how that precedent plays out is something the industry will be watching very closely.<\/p>\n In his policy essay published just days before the shutdown, Amodei opened with an analogy for the pacing mismatch between AI and regulation. \u201cIn one of the side plots to The Lord of the Rings, two of the Hobbits attempt to rouse Treebeard \u2014 a wise but ponderous sentient tree \u2014 to defend his forest from an army that is cutting it down. The problem is that Treebeard operates at a very different speed than the Hobbits.\u201d<\/p>\n \u201cThe intersection of AI and our political institutions feels a bit like the Hobbits and Treebeard,\u201d Amodei wrote.<\/p>\n This week, Treebeard moved fast.<\/p>\n<\/section>\n<\/div>\nWhat happened?<\/h2>\n
The Big \u2018Jailbreak\u2019<\/h2>\n
Amazon\u2019s Fingerprints<\/h2>\n
Equally Complex Backdrop<\/h2>\n