In brief: Anthropic offers Fable 5, a Mythos variant with safety filters for public use, while Project Glasswing participants gain access to less restricted Claude Mythos 5, accompanied by new federal rules controlling frontier AI models.
Anthropic has unveiled Fable 5, a Mythos-class model with safety measures against misuse for cyberattacks and bioweapons instruction. Simultaneously, Claude Mythos 5 has been released to Project Glasswing participants, while the U.S. government tightens controls on powerful AI models.
Anthropic announced two new models on Tuesday: Fable 5 is built on the same core architecture as the Mythos model but includes built-in safeguards. Questions about cybersecurity and biology are routed through the less powerful Claude Opus 4.8 model to make it more difficult for the model to be misused for planning cyberattacks or creating bioweapons instructions. The company describes Fable 5 as “the world’s most capable model for cybersecurity abilities.”
In parallel, Claude Mythos 5 was released as an upgrade to the previous Mythos Preview, which until now was available only to a small group of government agencies and private companies participating in Project Glasswing. This program enables cyber defenders in critical infrastructure to use Mythos to identify and remediate security vulnerabilities before the model becomes more widely available. Project Glasswing participants face fewer usage restrictions than Fable 5 users.
The release comes against a backdrop of increased regulatory requirements. President Trump signed an executive order requiring AI developers to submit certain new models to voluntary federal review 30 days before public release. Early tests show that models such as Mythos and OpenAI’s GPT-5.5-Cyber can find and exploit thousands of security vulnerabilities in networks. A bipartisan congressional proposal to control AI technologies was introduced on Capitol Hill in parallel.
Anthropic told POLITICO that the company has “regularly coordinated with the Administration on Mythos Preview as well as these new models” and feels committed to “close collaboration with the U.S. government as AI changes the cybersecurity landscape.” On whether Fable 5 was submitted to the federal government for review, the company did not comment explicitly.
Source: www.politico.com · Published June 9, 2026
Lumi AI News — AI-assisted curation pursuant to Article 50 EU AI Act. Paraphrase and classification by Lumi News Pipeline v1.6.5.