Anthropic restores global access to Fable 5

Anthropic has restored global access to Claude Fable 5 as of July 1, 2026. The model had been suspended on June 12 following a US government export control directive limiting its use to US nationals only; since real-time nationality verification was not feasible, access was disabled for all customers. On June 30 the controls were lifted and the model is available again.

The issue originated from a jailbreak technique found by Amazon researchers, which led Fable 5 to identify software vulnerabilities and, in one case, to demonstrate how to exploit them. Anthropic's testing confirmed, however, that every model examined — including GPT-5.5 and Claude Haiku 4.5 — could produce the same demonstration, with no offensive capabilities unique to Fable 5.

In response, Anthropic trained an improved safety classifier that blocks the problematic behavior in over 99% of cases, neutralizing the technique without exposing any unique capabilities. Together with Amazon, Microsoft, and Google, the company is also defining a shared framework to assess jailbreak severity across four criteria: the magnitude and breadth of capability gain, ease of weaponization, and discoverability of the technique.

On Pro, Max, Team, and select Enterprise plans, Fable 5 is included up to 50% of weekly usage limits through July 7, after which it remains available via usage credits. As an Anthropic partner, SETEK tracks the evolution of Claude models and their safeguards to deliver reliable, compliant AI solutions to enterprises and Public Administration.