Claude Fable 5 and Mythos 5 were temporarily suspended in June 2026, but that suspension is no longer in effect. Anthropic disabled access on June 12 after new US export controls took immediate effect. The company then announced on June 30 that the restrictions had been lifted and restored global access from July 1.
That timeline matters because widely shared social posts often describe the initial shutdown accurately while omitting the later restoration. Here is what happened, what the two models are designed to do, and what users should understand today.
What are Claude Fable 5 and Claude Mythos 5?
Anthropic introduced both models on June 9, 2026. They belong to the same model family but serve different audiences. Fable 5 is the broadly available version and includes stronger safeguards for sensitive areas. Mythos 5 is reserved for a limited group of trusted organisations, particularly those working in cybersecurity.
Anthropic described Fable 5 as its most capable generally available model at launch, highlighting software engineering, research, vision and long-running tasks. Mythos 5 retains more extensive vulnerability-research capabilities, which is why access to it remains tightly controlled.
Why did Anthropic switch them off on June 12?
According to Anthropic, the US government placed both models under export controls covering foreign nationals inside and outside the United States. The measure took effect immediately.
Anthropic said it had no reliable way to verify every user’s nationality in real time. Rather than risk breaching the directive, it suspended both Fable 5 and Mythos 5 for all users. The interruption therefore affected American users as well as customers abroad.
Is the suspension still active?
No. On June 30, Anthropic said the export controls affecting Fable 5 and Mythos 5 had been lifted. On July 1, the company confirmed that access had been restored.
Fable 5 returned to the relevant Claude products globally. Mythos 5 still follows Anthropic’s restricted-access programme, so restoration did not turn it into an unrestricted consumer chatbot.
Did the models discover “thousands of unknown vulnerabilities”?
Anthropic has reported advanced cybersecurity and vulnerability-discovery capabilities for the Mythos family. However, a dramatic number repeated in a social-media post should not be treated as independently verified without its testing methodology.
Results depend on the software examined, the definition of a vulnerability, the test environment and human validation. It is important to separate Anthropic’s own technical claims, independent evaluations and simplified viral summaries. We therefore do not present a precise number as established fact without documentation that supports it.
What the episode means for users
- A post published on June 12 or 13 may have been accurate then but is outdated today.
- Fable 5 returned to global availability from July 1.
- Mythos 5 remains a specialised, restricted-access model.
- Availability through cloud providers may follow separate technical schedules.
- An export-control restriction does not automatically mean that a model was hacked or compromised.
Verified timeline
- June 9, 2026: Anthropic announces Claude Fable 5 and Claude Mythos 5.
- June 12: the company suspends both models following new US export controls.
- June 30: Anthropic announces that the controls have been lifted.
- July 1: access to both models is restored under their respective availability rules.
Sources and verification
This article prioritises Anthropic’s official announcements: the model launch page and the restoration update. The government directive was also reported by the Associated Press.
Last checked: August 3, 2026. This article will be updated if access conditions change.