Version of AI tool 'too powerful for public' released to public

Claude Fable 5 is a version of Anthropic's Claude Mythos, an AI program which caused a stir among technology, finance, and government leaders.

Anthropic has announced the public release of its Fable 5 model, the first Mythos-class model made generally available, with safeguards to prevent misuse in areas like cybersecurity.

Anthropic has launched Claude Fable 5, the first publicly available version of its Mythos model, with hard safety limits in high-risk areas like cybersecurity and biology.

Anthropic has launched Claude Fable 5, a Mythos-class model that exceeds the capabilities of any model it has made generally available, with safeguards to prevent misuse in areas like cybersecurity.

Anthropic’s Claude Fable 5 AI model is now publicly available through the Claude desktop app, with major improvements in coding, spatial reasoning, and cybersecurity capabilities, but access requires usage credits after June 23.

Anthropic's Mythos Preview demonstrated unprecedented capability by completing a 32-stage network attack simulation, marking a significant escalation in real-world AI-assisted cyberattacks.

Anthropic launched Claude Fable 5 on June 9, 2026, calling it the most capable model it has ever released to the general public. Alongside it, the company released Claude Mythos 5 for a small group of cybersecurity defenders and infrastructure providers — the same underlying model as Fable 5, but with some safeguards lifted.

As of June 2026, Anthropic and OpenAI have emerged as central figures in cybersecurity by controlling access to their highly advanced AI models through selective access programs. OpenAI has already implemented a 'trusted-access' initiative, and Anthropic is planning to launch a similar program for its Mythos class, particularly the newly announced Fable 5 model now available to the public. These programs allow these AI labs to determine which cybersecurity professionals and organizations can use their most cyber-capable tools, effectively positioning them as gatekeepers of cutting-edge defensive—and potentially offensive—cyber technologies. The strategy allows the companies to retain control over sensitive capabilities to prevent misuse, while still monetizing their powerful models as they eye public markets. The dual-access model provides exclusive advantages to those granted trusted access, potentially enabling them to identify vulnerabilities, conduct advanced research, and develop superior cybersecurity products. This raises broader concerns about fairness and reliance on private AI firms in critical cybersecurity domains.

Anthropic's new AI model, Mythos Preview, has demonstrated the capability to exploit newly disclosed software vulnerabilities within hours, a significant acceleration compared to the weeks traditionally required. In an exclusive report, Anthropic revealed that its frontier red team used Mythos to successfully develop a proof-of-concept exploit for a Microsoft Windows kernel flaw just 31 minutes after disclosure. The test also included vulnerabilities in Mozilla Firefox. This showcases the potential for AI to not only identify new bugs but also quickly weaponize known ones, raising concerns about cybersecurity. While AI's bug-finding abilities have been in the spotlight, this development shifts attention to its rapid exploitation capabilities. Mythos is not alone—other open-source AI models and competitors like GPT-5.5-Cyber are demonstrating similar performance. This comes as the U.S. government, under the Trump administration, begins enforcing a new AI security executive order to monitor risks posed by advanced AI. The move aligns with broader industry shifts, including Cisco adapting its vulnerability disclosure processes to address the AI era.

Anthropic has made its most powerful AI model, Claude Fable 5, available to the general public, implementing guardrails to prevent misuse in high-risk areas like cybersecurity and biology.

Anthropic's Claude Fable 5, a Mythos-class model, is now publicly available, offering state-of-the-art performance across various benchmarks with safeguards to prevent misuse in sensitive areas.

Anthropic has made a new version of its Claude Mythos AI model, Fable 5, available to the general public, with safeguards to prevent misuse in sensitive areas like cybersecurity and biological capabilities.

Anthropic's Claude Fable 5, a version of the previously restricted Claude Mythos AI program, has been released to the public, raising concerns among technology, finance, and government leaders.

Anthropic has made Claude Fable 5, a safer version of its powerful AI model Mythos, available to the public, with usage credits required after June 22.

Anthropic's new AI model, Claude Fable 5, has been released to the public, raising concerns about its potential misuse due to its powerful capabilities.

Anthropic has launched Claude Fable 5, a new AI model built on the same foundation as the previously restricted Claude Mythos 5, a system originally deemed too powerful for public release due to its advanced cybersecurity capabilities.

A new report from the Financial Times claims the NSA is utilizing Anthropic’s cybersecurity-focused AI model, Mythos, for offensive cyber operations, despite the Department of Defense (DOD) having previously labeled the company a 'supply chain risk' and banning it from providing AI services. Allegedly, six Anthropic engineers are embedded at the NSA to help customize and deploy the Mythos model for specialized cyber attack applications, targeting potential adversaries like China and Iran. This controversial development comes amid ongoing lawsuits between Anthropic and the DOD, sparked after Anthropic refused contract terms that would have required the removal of AI safety guardrails for government use. The $200 million contract was then terminated, and Anthropic was sanctioned similarly to Chinese tech firms Huawei and ZTE. While the DOD partnered with OpenAI instead, critics allege the decision was politically motivated. The story also surfaces as the Trump administration issues an executive order mandating that leading AI models be submitted to the government for early access. Mythos, still in early access, was recently offered to 150 organizations worldwide. Despite its legal troubles, Anthropic’s AI models are reportedly advancing rapidly, raising concerns about the pace of AI development.