Anthropic Prepares Claude Mythos for Wider Release

Anthropic is preparing a commercially versioned release of its most capable frontier model, reported as Mythos 1 with the model identifier claude-mythos-1-preview. TestingCatalog and BleepingComputer both report traces of the claude-mythos-1-preview string in Anthropic product code and UI text for Claude Code and Claude Security. BleepingComputer reports that some users briefly saw a toggle to enable Mythos in the public Claude Code interface before the toggle was removed. ITSecurityNews and related coverage point to an Anthropic Project Glasswing update that used the unreleased Mythos preview during vulnerability discovery efforts and, according to that reporting, identified more than 10,000 high- and critical-severity zero-day vulnerabilities.

BleepingComputer reports Anthropic described the Mythos preview as showing "major improvements in code reasoning and autonomy" above the company's then-flagship Opus 4.7, and that Anthropic found the model could generate professional-grade offensive code during internal testing. TestingCatalog reproduces a Project Glasswing line stating: "And in the near future, once we've developed the far stronger safeguards we need, we look forward to making Mythos-class models available through a general release." The coverage also notes UI and source-code strings added to product surfaces and a refreshed Claude Security dashboard intended to surface discovered vulnerabilities, historical charts, and triage results for enterprise customers.

Editorial analysis: Frontier models that materially raise autonomous code-generation and vulnerability-discovery capabilities create a pronounced security-versus-availability tradeoff. Industry observers have previously documented that labs deploying high-capability models commonly gate access, iterate on guardrails, and coordinate with partners before widening availability. When guardrails and monitoring dashboards are mature, vendors typically expand enterprise access first and then broader tiers over time; the signals in the product UI and Project Glasswing updates fit that familiar pattern.

Editorial analysis: For security teams and ML engineers, a publicly available Mythos-class model would change tooling and threat models. On the defensive side, access to high-capability vulnerability-finding models can accelerate triage and automated patching workflows. On the offensive side, the same capabilities lower the barrier for automated exploit generation if access controls or usage constraints are insufficient. The dual-use nature means operational controls, provenance, and monitoring are likely to be the key determinants of net impact as availability changes.

Editorial analysis: Observers should track four indicators:

• •official product release notes or developer docs that list claude-mythos-1-preview or Mythos 1
• •whether Mythos appears in paid tiers or remains enterprise-only
• •changes to the Claude Security dashboard and what telemetry it exposes
• •further Project Glasswing publications describing methodology and safeguards. Also watch cloud vendor listings and SDK/API references for concrete rollout signals

BleepingComputer reproduces Anthropic warnings quoted on risk: "The advantage will belong to the side that can get the most out of these tools," and "In the short term, this could be attackers, if frontier labs aren't careful about how they release these models. In the long term, we expect it will be defenders who will more efficiently direct resources and use these models to fix bugs before new code ever ships." TestingCatalog and ITSecurityNews report Project Glasswing briefings that connect the Mythos preview to collaborative vulnerability discovery work.

Overall, the reporting presents converging signals of a product-phase change for a frontier, dual-use model. Sources document UI traces, dashboard work, and an explicit Project Glasswing linkage; none of the scraped reports cite a formal general-availability announcement from Anthropic at the time of publication.