Anthropic's Claude Fable 5 returns after Trump admin lifts export controls

After weeks of tense negotiations with the Trump administration, Anthropic has secured the return of its consumer-facing Claude Fable 5 model. The company announced Tuesday evening that the Department of Commerce has lifted export controls on both Fable 5 and its underlying Mythos 5 model. Access will begin restoring Wednesday on Claude platforms, with AWS, Google Cloud, and Microsoft Foundry to follow. The models were sidelined in early June after the government issued an export control directive banning foreign nationals—including Anthropic's own employees—from using them, citing a jailbreak technique flagged by Amazon researchers.

To address the security vulnerability, Anthropic trained an improved safety classifier that blocks the specific jailbreak technique in over 99% of cases. Users whose requests are blocked will receive a notification and have their query redirected to Opus 4.8. The company also outlined a new framework for working with the government, including pre-release government access and evaluation for models relevant to national security, rapid information sharing when jailbreaks are detected, and a voluntary shared testing process with other leading AI labs. This comes shortly after OpenAI's GPT-5.6 debuted under similar staggered rollout restrictions.