Claude Fable 5 Still Down on Day Nine - Anthropic's Deadline Has Passed
Anthropic's managing director promised the models would return "within days" on June 18. Nine days after the US export ban, nothing has changed.
Claude Fable 5 has now been offline for nine days. Anthropic's managing director of international, Chris Ciauri, told reporters in Seoul on June 18 that the company was "very confident" the models would return "within days." Access has not returned. API calls to claude-fable-5 still return errors, and Anthropic has issued no public update since Ciauri's statement.
Commerce Department Ordered the Ban - Anthropic Could Not Comply Selectively
On June 12, the US Commerce Department issued an export control directive ordering Anthropic to suspend access to Claude Fable 5 and Claude Mythos 5 for any foreign national - inside or outside the United States. Anthropic cannot verify user nationality in real time at the API level. So rather than attempt partial compliance and risk violating the directive, Anthropic switched off both models globally, cutting off US users alongside everyone else.
Amazon CEO Andy Jassy triggered the action, disclosing to Treasury and Commerce officials that Amazon researchers had found prompts capable of bypassing Fable 5's safety filters to extract software vulnerability data. Anthropic called the directive a "misunderstanding" and said it disagreed with the government's interpretation.
We are very confident that in the coming days, the models will become available again.
Chris Ciauri, Managing Director of International, Anthropic - June 18, 2026
Over 100 Security Experts Dispute the Government's Jailbreak Claim
An open letter organized by Alex Stamos - former chief of security at Facebook - started with 76 signatories and has grown past 100. Names from Nvidia, Google, Adobe, Zoom, and Sophos all appear on the list. Stamos and the co-signatories addressed it to Commerce Secretary Howard Lutnick and National Cyber Director Sean Cairncross.
Security researcher Katie Moussouris described the alleged jailbreak as rephrasing a prompt as "fix this code" - a technique used daily in defensive security work, not an exploit. Signatories argue the same capability exists in GPT-5.5, Claude Opus 4.8, and China's Kimi 2.7. All three models remain fully available. Pulling Fable 5 removed the best tool from the defenders who needed it most, without meaningfully reducing access for anyone with bad intent - which is probably the sharpest argument in the letter.
For developers who relied on Fable 5 for security audits and vulnerability research, Anthropic's original suspension announcement offered no timeline and no migration path. Opus 4.8 is still available, but teams built around Fable 5's extended reasoning profile are bridging an unexpected gap with no confirmed end date.
Kalshi Traders Give 58% Odds on Restoration Before July 1
Prediction market traders on Kalshi price a 58% chance that Fable 5 returns before July 1, rising to 74% by mid-July. Those numbers suggest the market reads this as temporary - but not close to resolution.
Four open-weights models, including Llama and Mistral variants, moved to fill the gap within days of the ban, according to The New Stack. Enterprise customers locked into Anthropic's API bore the heaviest cost. Developers with model flexibility found alternatives within 48 hours.
Anthropic has not published a revised timeline since Ciauri's June 18 statement. Commerce has not commented publicly on the open letter.
