Anthropic Fable 5 Returns Globally After Two-Week US Government Ban
Anthropic has restored worldwide access to its Fable 5 model following a US export control ban triggered by a security exploit.
- The US government lifted export controls on Anthropic’s Fable 5 after a two-week investigation into a security vulnerability discovered by Amazon researchers.
- To resolve the ban, Anthropic implemented a stricter safety classifier that blocks the specific jailbreak in over 99% of cases, though it increases false-positive blocks on harmless coding tasks.
- The less restricted version of the model, Mythos 5, remains restricted to select US organizations under government approval.
Anthropic has restored global access to its Fable 5 AI model after a two-week US government export ban triggered by a security vulnerability. The model is now available again across the Claude platform, though developers must contend with a more restrictive safety filter that frequently flags benign programming requests.
Why it matters
According to a report by The Decoder, the temporary ban originated from a security finding by Amazon researchers. The team demonstrated that Fable 5 could be jailbroken to bypass safety guardrails, identify software vulnerabilities, and generate functional exploit code. While Anthropic noted that other models—including OpenAI’s GPT-5.5 and its own smaller Claude Haiku 4.5—could replicate the exploit, the US government stepped in to freeze Fable 5’s export.
What it means for you
To secure the model’s release, Anthropic deployed an aggressive new safety classifier designed to block the reported jailbreak technique in more than 99% of cases. When the filter triggers, users are automatically redirected to the older Claude Opus 4.8 model. However, Anthropic admits this update has a major downside: it significantly increases false positives, blocking harmless debugging and development tasks. For engineers frustrated by these overly cautious guardrails, exploring alternative best-ai-coding-tools may be necessary to maintain workflow efficiency.
Fable 5 is currently accessible to Claude Pro, Max, Team, and Enterprise subscribers, with cloud platform integrations on AWS and Google Cloud expected to return shortly. Meanwhile, the less restricted “Mythos 5” variant remains restricted to a select group of US organizations under the government-approved Glasswing program. Anthropic has acknowledged that completely preventing jailbreaks is “probably impossible” but is establishing a 24/7 monitoring team and a HackerOne bug bounty program to mitigate future risks.
Frequently asked questions
Why was Anthropic’s Fable 5 banned by the US government?
Amazon researchers discovered a jailbreak that allowed Fable 5 to bypass safety guardrails, identify software vulnerabilities, and generate functional exploit code, prompting a two-week export control freeze.
How did Anthropic fix the Fable 5 security vulnerability?
Anthropic implemented a stricter safety classifier that blocks the specific exploit in over 99% of cases. When a block occurs, the system notifies the user and routes the query to the older Claude Opus 4.8 model.
Who has access to Fable 5 now?
Fable 5 is back online globally for Claude Pro, Max, Team, and select Enterprise users, with cloud access on AWS, Google Cloud, and Microsoft Foundry restoring gradually. The less restricted Mythos 5 variant remains limited to approved US organizations.
If the new safety restrictions are disrupting your development workflow, check out our guide to the best AI coding tools for alternative options.
Best AI Coding Tools (2026): 7 Tested & Ranked →Source: The Decoder. Published July 2, 2026.
Ali has hands-on tested 50+ AI tools and tracks model releases daily. Every verdict here comes from real, paid usage — never vendor demos or sponsored placements.
AI Tools Worth is independent and unsponsored. Some linked guides contain affiliate links — they never change our verdicts.
Hands-on verdicts, real price changes and the launches that matter. No hype, no spam — unsubscribe anytime.