Anthropic Halts Access to Fable 5 and Mythos 5 Amid U.S. Government Directive Anthropic announced on Friday that it has disabled access to its Fable 5 and Mythos 5 artificial intelligence models to comply with a U.S. government directive citing "national security authorities." The company stated it received the order at 5:21 p.m. ET, instructing it to suspend all access to the models "by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees." The abrupt decision affected all customers, though the company emphasized that its other models remain unaffected. The move follows the recent release of Fable 5 and Mythos 5, two advanced AI models that Anthropic had positioned as state-of-the-art across multiple industry benchmarks. Fable 5, in particular, marked the first time the company publicly released such a sophisticated offering, thanks to new safeguards designed to block responses in high-risk areas. The models were built on the foundation of the Claude Mythos Preview, which garnered attention in April for its cybersecurity capabilities. However, Anthropic stated it did not plan to make the model generally available, instead limiting its rollout to a select group of companies through a cybersecurity initiative called Project Glasswing. In its statement, Anthropic noted the government did not provide specific details about its national security concerns. The company apologized to customers for the disruption, reiterating its belief that the government should have the ability to block unsafe deployments through a transparent, fair, and technically grounded process. "This action does not adhere to those principles," the company said, highlighting its frustration with the lack of clarity in the directive.#anthropic #department_of_defense #project_glasswing #mythos_5 #fable_5
Claude Opus 4.8 Released with Enhanced Features and Performance The latest iteration of the Claude Opus series, version 4.8, has been launched with a range of improvements aimed at enhancing collaboration, efficiency, and reliability. This update builds on the foundation of its predecessor, Opus 4.7, and introduces new capabilities designed to better serve users in coding, reasoning, and practical knowledge tasks. The release coincides with several key features, including dynamic workflows for Claude Code and a faster, more cost-effective fast mode. These updates are now available at the same pricing as previous versions, with no changes to the standard rate structure. Opus 4.8 introduces a new feature that allows users to control the amount of effort the model dedicates to a task. This flexibility enables users to balance quality and speed based on their specific needs. For coding tasks, the default setting is set to "high effort," which the developers claim offers the best overall balance between performance and user experience. Users can also opt for "extra" or "max" effort levels, which allocate more computational resources to achieve higher accuracy, particularly for complex or long-running workflows. To accommodate these settings, rate limits in Claude Code have been increased, ensuring that users can select the most appropriate option for their projects. A significant focus of the update is on improving the model's reliability and judgment, especially in agentic tasks where the model must make decisions or take initiative. Early testers have reported that Opus 4.8 is more honest and less prone to making unsupported claims. This improvement is attributed to enhanced training processes that prioritize transparency and accuracy.#anthropic #claude_code #claude_opus #claude_mythos #project_glasswing
