When will Fable 5 be back online? Polymarket prediction market data: 75% chance of unblocking in July

The U.S. Department of Commerce has imposed export controls on the grounds of “national security,” prohibiting foreigners from accessing Anthropic’s newly released Claude Fable 5. Currently, Polymarket betting markets indicate a 75% chance that access for U.S. users will be restored before July 1.
(Background: The most powerful AI from Anthropic is restricted from export by the U.S.! Fable 5, Mythos 5 are cut off worldwide)
(Additional context: Can Claude Fable 5 be unblocked? Presidential adviser: Fixing one issue can lift the restrictions, letting the “fable” go back online.)

In just three days, Claude Fable 5 went from public launch to being forcibly taken down at the request of the U.S. government—only three days. This morning (14), presidential technology adviser David Sacks laid out a path to get it unblocked: as long as the jailbreak vulnerability is patched, it can “bring Fable back online.”

So quickly, Polymarket appeared just a few hours ago with a new betting market: “When will Claude Fable 5 be restored for U.S. users?” As of the time of publication, the total trading volume of this prediction market is about $400k. The market pricing for three expiration dates is as follows:

• 12% chance of restoration before June 15
• 56% chance of restoration before June 22
• 75% chance of restoration before July 1.

Jailbreak vulnerabilities and selective enforcement: Why Anthropic isn’t convinced

According to an earlier report by DeepZone, David Sacks, co-chair of the U.S. Presidential Technology Advisory Committee, said the reason for the takedown was that a highly trusted partner—trusted by both Anthropic and the U.S. government—found a jailbreak method that could break through the guardrails while testing Fable (i.e., inducing the model to bypass built-in restrictions).

However, Anthropic’s position is that this jailbreak is “narrow and not universal,” and that the same vulnerability also exists in other publicly available models such as OpenAI’s GPT-5.5, but those models have not been subjected to export controls. Anthropic believes that if the standard of “recalling an entire commercial model due to a single narrow jailbreak” is applied across the industry, it effectively freezes the deployment of all frontier models.