Anthropic's Fable 5 model expected to return next week after 15-day government shutdown
The Trump administration is close to allowing Anthropic to restore access to its Fable 5 model, which has been offline for 15 days due to national security concerns. Insiders expect restrictions could be lifted as soon as next week, though Pentagon and NSA approval is still required.
Claude Fable 5 — Quick Specs
Anthropic's Fable 5 Expected to Return After 15-Day Government Shutdown
The Trump administration is close to allowing Anthropic to restore access to its Fable 5 model, which has been offline for 15 days due to government security concerns, according to sources close to the situation.
Insiders expect the administration's limits on Fable 5 could be lifted as soon as this coming week. A second source said conversations are expected to continue over the weekend, and Anthropic expects to restore Fable access soon.
Background on the Shutdown
Fable 5 was pulled offline on June 12, just three days after launch, following a standoff between the Trump administration and Anthropic. The shutdown marked an unprecedented moment for the AI industry — a top-tier model, already in users' hands, removed due to government intervention.
The model had been available at no extra cost on several paid Claude subscription plans through June 22, giving users a brief window to test its capabilities before access vanished.
Signs of De-escalation
In a sign of thawing relations, the Commerce Department on Friday allowed Anthropic to restore limited access to Mythos 5, the company's strongest cybersecurity model. Mythos 5 has guardrails to deter its use in cyberattacks or biological terror and has never been freely available.
Commerce Secretary Howard Lutnick said in a Friday letter to Anthropic that the company "has worked with the U.S. government to address risks associated with" both models and "has committed to work with the U.S. government on protocols and standards and releases."
Fable 5's Capabilities
According to Anthropic, Fable 5 was "the most capable model ever released to the public." The Every newsletter called it "the best coding model in the world" before it was pulled.
In early testing highlighted by Anthropic, payments company Stripe used Fable 5 to overhaul a 50-million-line codebase in a single day — a job that would have taken its engineers more than two months by hand. Users praised the model's deep thinking and sophisticated coding capabilities.
When access vanished on June 12, developers found automated work frozen mid-task, and companies raced to swap in rival models, including cheaper Chinese alternatives.
Outstanding Approvals
The Pentagon and National Security Agency still have to give Fable 5 the green light, so the outcome remains unpredictable. However, other government agencies have determined Fable 5 can safely return.
Both Commerce Secretary Lutnick and Treasury Secretary Scott Bessent have helped defuse the fight between the administration and Anthropic, according to sources. This represents a significant change from Defense Secretary Pete Hegseth's earlier statement designating Anthropic a "Supply-Chain Risk to National Security."
Uncertain Return Terms
It's not yet clear whether Anthropic subscribers will get back the free access to Fable they were promised — or whether it returns locked behind additional fees or identity checks.
Industry Push for Clear Process
Both Anthropic and OpenAI are pushing the administration to codify a process for reviewing new models, as envisioned by President Trump in a June 2 executive order that set up a framework for voluntary government vetting of the most powerful new AI models.
When Anthropic suspended access to Fable 5 and Mythos 5 on June 12, the company called for "a statutory process that is transparent, fair, clear, and grounded in technical facts," saying the administration's decision "does not adhere to those principles."
When OpenAI was allowed to begin a limited preview of GPT-5.6 on Friday, the company stated: "We don't believe this kind of government access process should become the long-term default. It keeps the best tools from users, developers, enterprises, cyber defenders, and global partners who need them."
What This Means
This situation marks the first time the U.S. government has forced a commercial AI model offline after public release, setting a precedent that could reshape how frontier AI models are deployed. The lack of a clear, codified review process has created uncertainty for AI companies about what triggers government intervention and how long reviews will take. If Fable 5 returns with restricted access or additional fees, it could signal a new tier of government-approved AI models, fundamentally changing the economics and availability of frontier AI capabilities.
Related Articles
Anthropic's Claude Fable 5.1 Reportedly Solves 1653 Royalist Cipher in 44 Minutes
According to testing firm Vals AI, Anthropic's Claude Fable 5.1 independently identified and solved the 'Cyphral Distich,' a 1653 numeric cipher by Sir Thomas Urquhart that had defeated other frontier models. The AI decoded a hidden pro-royalist message by mapping each number to a word in Urquhart's original text.
Anthropic Brings Background Computer Use to Claude Code and Cowork on Mac
Anthropic has enabled background computer use for Claude Code and Claude Cowork on macOS, available to Pro and Max subscribers. The feature lets Claude click, type, and open apps on a Mac without taking over the user's active cursor, following a similar launch by OpenAI's ChatGPT earlier in 2026.
OpenAI's GPT-6 Astra Cuts Hallucinations, But Indirect Prompt Injection Attacks Still Succeed 8.5% of the Time
OpenAI's new GPT-6 Astra model shows major improvements in hallucination rates and jailbreak resistance over predecessor GPT-5.6 Sol, according to OpenAI's system card. However, indirect prompt injection attacks hidden in documents still succeed 8.5% of the time in external testing by Gray Swan, down from 27% but still above rival Claude Opus 5's 4.8% rate.
OpenAI Ships GPT-6 Astra, But Executives Admit They Can't Fully Monitor What It's Thinking
OpenAI released GPT-6 Astra on Thursday, a model president Greg Brockman says could mark the start of AGI. But the model writes out its reasoning less often than prior versions, and OpenAI's chief scientist says monitoring AI thought processes will keep getting harder.
Comments
Loading...