News & Updates, AI Agents, AI Assistants

US Government Forces Anthropic to Pull Claude Fable 5 Globally

Anthropic suspended access to Claude Fable 5 and Mythos 5 for all users worldwide on June 12 after the US government issued an export control directive citing a jailbreak. Anthropic is contesting the order while complying.

3 min read
US Government Forces Anthropic to Pull Claude Fable 5 Globally

Anthropic

US Government Forces Anthropic to Pull Claude Fable 5 Globally

Anthropic has suspended access to Claude Fable 5 and Claude Mythos 5 for all users worldwide following a US government export control directive received at 5:21 PM ET on June 12—just three days after the models launched on June 9.

The government's directive, issued under national security authorities, orders Anthropic to block all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. Because Anthropic cannot reliably identify users' citizenship status in real time, the company complied by disabling access globally for every customer. All other Claude models—Opus 4.8, Sonnet 4.6, Haiku 4.5, and earlier versions—remain unaffected and operational.

Anthropic published a full statement on its website disputing the basis for the order.

What the Government Said

The directive cited national security concerns linked to what the government described as a method for bypassing Fable 5's safeguards—a so-called jailbreak. Anthropic says it received only verbal evidence of the alleged technique, which it characterizes as a narrow, non-universal jailbreak that amounts to asking the model to read a codebase and fix software vulnerabilities.

In its statement, Anthropic wrote: "We have reviewed a report that we believe is the basis of the government's directive and validated that the level of capability displayed there is widely available from other models (including OpenAI's GPT-5.5), and is used every day by the defenders who keep systems safe."

The government has not publicly shared technical details of the alleged jailbreak, and Anthropic says the finding represents "minor findings that provide no Mythos-specific uplift" beyond what already exists in models deployed across the industry.

Developer Impact

For developers who migrated to Fable 5 or Mythos 5 after their June 9 general availability launch, the shutdown is an immediate production incident. Any application, pipeline, or Claude Code workflow that specifies either model as its backend stopped working as of Thursday evening. Developers relying on Claude Code's ultracode mode—which routes to Fable 5—are also affected.

Fallback paths are available. Anthropic confirmed that all other models remain operational, so downgrading to Claude Opus 4.8 or Sonnet 4.6 is the immediate mitigation. Developers using the Claude API can swap the model parameter. Claude Code users can switch the active model via the effort menu in the IDE.

The incident exposes a new category of operational risk: government-ordered model recall. Unlike infrastructure outages, which typically resolve within hours, this shutdown has no confirmed restoration timeline. Anthropic said it is "working to restore access as soon as possible" and believes the suspension is based on a misunderstanding, but the government has not indicated any plans to rescind the directive.

Anthropic's Position

Anthropic publicly disagreed with the order while complying with it. The company argued that its defense-in-depth approach to Fable 5's safety architecture—extensive red-teaming, 30-day customer data retention for monitoring, and strong anti-jailbreak safeguards—already addresses the government's concern without a global shutdown.

"We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people," Anthropic wrote in its official statement. "If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers."

The company also noted that no testers have found a universal jailbreak—one that broadly bypasses Fable 5's safeguards across a wide range of tasks—and stated that perfect jailbreak resistance is not currently achievable by any model provider in the industry.

Anthropic said it plans to publish additional technical details about the alleged jailbreak within 24 hours of the suspension.

What's Unconfirmed

Anthropic has not disclosed when access to Fable 5 or Mythos 5 will be restored. The government has not released its technical evidence publicly. It is also unclear whether the suspension extends to Fable 5 and Mythos 5 access via Amazon Bedrock or Google Vertex AI, or applies only to the direct Anthropic API and claude.ai products. Developers using third-party hosting should verify model availability with their cloud provider directly.

Share:

Other Latest News

SpaceX Lists Today, Starting 30-Day Clock on $60B Cursor Buyout
Code Editors, News & Updates, Industry Analysis

SpaceX Lists Today, Starting 30-Day Clock on $60B Cursor Buyout

SpaceX begins trading on Nasdaq today under SPCX, triggering the 30-day countdown before the company moves to close its $60 billion acquisition of AI coding platform Cursor.

Jun 13, 2026
Meta Cuts Off Manus Access as $2B Acquisition Unwinds
AI Agents, News & Updates, Industry Analysis

Meta Cuts Off Manus Access as $2B Acquisition Unwinds

Meta is severing Manus from its internal systems and barring employees from using the platform, operationally unwinding the $2 billion acquisition after Beijing ordered the deal reversed in April.

Jun 13, 2026
Windsurf Becomes Devin Desktop With New Agent Command Center
AI Agents, News & Updates, Code Editors

Windsurf Becomes Devin Desktop With New Agent Command Center

Cognition has relaunched Windsurf as Devin Desktop, a full IDE with an Agent Command Center and open ACP protocol support that lets Codex, Claude Agent, OpenCode, and custom in-house agents run side by side in a shared Kanban workspace.

Jun 3, 2026
OpenAI Codex Lands on Amazon Bedrock for 5 Million Weekly Users
AI Agents, News & Updates, Code Editors

OpenAI Codex Lands on Amazon Bedrock for 5 Million Weekly Users

OpenAI's AI coding agent Codex is now generally available on Amazon Bedrock, giving developers at AWS-native organizations direct access to Codex through existing cloud infrastructure, procurement, and compliance workflows.

Jun 3, 2026
Anthropic Ships Claude Opus 4.8 With Dynamic Workflows and 3x Cheaper Fast Mode
AI Agents, News & Updates, Code Editors

Anthropic Ships Claude Opus 4.8 With Dynamic Workflows and 3x Cheaper Fast Mode

Anthropic releases Claude Opus 4.8 with improved agentic coding benchmarks, a 3x cheaper Fast Mode, and Dynamic Workflows that run hundreds of parallel subagents inside Claude Code.

May 29, 2026
MiniMax Revenue Doubles as M3 Promises 15x Long-Context Speed Boost
News & Updates, AI Assistants, Industry Analysis

MiniMax Revenue Doubles as M3 Promises 15x Long-Context Speed Boost

MiniMax annualized revenue doubled in two months, crossing 1M enterprise users. The company just teased M3's sparse attention architecture claiming 15.6x faster decoding at 1M-token contexts.

May 28, 2026
← Scroll for more →