Anthropic restores Claude Fable 5 and Mythos 5 as US lifts export controls

Anthropic, the AI safety company behind the Claude family of models, announced on July 1, 2026, that the U.S. Department of Commerce has lifted export controls on its two most advanced models: Claude Fable 5 and Claude Mythos 5. These models were previously disabled for international users in early 2026 over national security concerns, specifically around potential misuse of their capabilities for disinformation, cyberattacks, or weapons development. With the controls now removed, Anthropic is restoring full access to both models, allowing enterprises, developers, and researchers worldwide to again leverage their state-of-the-art reasoning, multimodal understanding, and autonomous agent features.

To secure this regulatory green light, Anthropic has committed to a new framework of proactive security risk detection, including real-time monitoring and mandatory reporting of any misuse or emergent dangerous behaviors. The company will also collaborate with U.S. government agencies—including the Commerce Department and national security offices—on standardized protocols for evaluating and deploying future models. This agreement could set a precedent for how other AI labs navigate export restrictions, balancing innovation with national security. For Anthropic, it means its flagship models are back in the hands of paying customers, a critical step for revenue and market share amid intensifying competition from OpenAI and Google DeepMind.