Anthropic has launched Claude Sonnet 5 and restored access to its Fable 5 and Mythos 5 frontier models after completing a federal export control review. The company had faced an 18-day operational pause following a government directive triggered by researchers demonstrating a method to bypass the models’ safety controls. Anthropic developed an updated automated classifier to address the vulnerability, allowing commercial rollout to resume across its platform and partner networks. Claude Sonnet 5 offers strong performance on autonomous agent tasks while maintaining safety standards comparable to or better than prior generations.
