SentinelOne says it will add Anthropic's Claude Mythos 5.1 to Wayfinder Frontier AI Services, a service in which its security experts use frontier AI models to find weaknesses in a customer's code and systems. The company announced the plan in a press release on October 29, describing it as an "upcoming addition". The release does not give a date for when the model will be in use, and it does not list prices.
The announcement shows one way Mythos 5.1 can reach customers. Anthropic does not sell Mythos 5.1 to the general public. It is the same model as Claude Fable 5.1, but with more permissive safeguards for cybersecurity and life sciences work, and Anthropic limits it to vetted individuals and organizations. For now, Anthropic says, that means a set of US organizations. A managed service like Wayfinder is one way that kind of model ends up working on companies' systems, with a security firm's staff in between.
What SentinelOne is adding
According to SentinelOne, Wayfinder Frontier AI Services is "a human + AI offering, backed by the latest frontier models from Anthropic, OpenAI and other frontier labs and run by an elite team of offensive and defensive security experts." With Mythos 5.1 added, the release says, "those experts will be able to apply Anthropic's most capable model class for cybersecurity to discover potential risks, then validate which findings are exploitable today."
The release lists three things customers get:
- Frontier-model discovery. The Wayfinder team uses frontier models to identify vulnerabilities and exposures in a customer's code and environment.
- Exploitability validation. SentinelOne's experts confirm which findings are "truly exploitable today", so remediation focuses on real risk rather than theoretical severity.
- Prioritized remediation guidance. The team delivers "a clear order of operations for fixes", so security teams know what to close first.
The Wayfinder Frontier AI Services page adds that findings can flow into SentinelOne's other Wayfinder services, including threat hunting, managed detection and response, and incident readiness and response. The same page says the service draws on SentinelOne telemetry from "tens of millions of endpoints", intelligence from SentinelLABS and Google Threat Intelligence, and frontier AI models including Claude, with "human judgement at every decision point."
Why SentinelOne says it needs this now
SentinelOne frames the move around what it calls a "tsunami of AI-fueled vulnerabilities and attacks that many have dubbed 'Black October'." The release argues that frontier models "can surface weaknesses and open the door to exploits faster than most teams can triage them," and that "a longer list does not help a defender."
Steve Stone, SentinelOne's Chief Customer Officer, said in the release: "Throughout October, we've seen a sharp increase in the volume and severity of vulnerability disclosures and cyberattacks, including PHP2SHELL, LiteLLM, eMudhra, and dozens more." He added: "Adding Anthropic's Claude Mythos 5.1 will help our analysts work at the pace of today's threats, so customers get validated findings and a clear order of operations for what to fix first."
Investing.com also reported the announcement, based on the press release.
What Claude Mythos 5.1 is
Anthropic introduced Claude Fable 5.1 and Claude Mythos 5.1 in September. Anthropic says the two "are the same model, but with different levels of safeguards." Fable 5.1 is generally available. Mythos 5.1 "is available only through our trusted access programs," and Anthropic says its safeguards "are specifically designed to support work in cybersecurity and the life sciences."
The difference matters for security work. Anthropic says Fable 5.1 can now be used to find software vulnerabilities, but its safeguards "still redirect several kinds of dual-use cybersecurity tasks" to Opus models, including penetration testing, exploit generation and binary-based vulnerability scanning. Mythos 5.1, by contrast, "offers more permissive safeguards for vetted individuals and organizations."
Anthropic also says that, with cybersecurity safeguards off, Mythos 5.1 "demonstrates the strongest cyber capabilities of any model we've released," while still falling "within the lower category of risk" in its Frontier Compliance Framework. Anthropic's own vulnerability scanner, Claude Security, is also powered by Mythos 5.1, according to the launch post.
How companies get access
On October 6, Anthropic expanded its Cyber Verification Program into three access tiers: Defense Access, Red Team Access and Specialized Access. Anthropic says each tier "includes access to our most capable models, including Claude Opus 5.5, Claude Sonnet 5.5, Claude Mythos 5.1, and new models moving forward." The tiers differ in what work is allowed. Red Team Access adds authorized penetration testing, and Specialized Access is reserved for a limited set of verified organizations, reviewed in collaboration with the US government.
Anthropic also requires data retention for organizations in the program "so that we can monitor for cyber misuse." Until its Enterprise Frontier Safeguards are available, organizations that already use Fable 5.1 or Mythos 5.1 with zero data retention can also use the program with zero data retention.
SentinelOne's release does not say which access tier it holds or how its use of Mythos 5.1 is set up with Anthropic.
Why it matters
For most companies, the practical question is not whether a model like Mythos 5.1 can find bugs, but how they would use it safely. Anthropic gates direct access through verification. SentinelOne is offering a second route: hire its team, which runs the models and decides which results are real before handing over a fix list. That puts a human review step between a powerful cyber model and the customer's systems.
Other labs are also offering models to security teams. Google said on October 28 that it will roll out Gemini 4 Carbon first for cyber defenders, and SentinelOne's release notes that Wayfinder already uses models from OpenAI and other labs as well. At the same time, misuse of AI coding tools remains a live issue, as in the case Symantec reported on October 29 of an unauthorized relay service that switched off Codex and Claude Code sandboxes. Anthropic has also published its own review of a Mythos 5 cyber test incident.
What to watch next: when SentinelOne says Mythos 5.1 is live in Wayfinder, and whether other security service firms announce similar access.
Sources
- SentinelOne: SentinelOne to Add Anthropic's Claude Mythos 5.1 to Wayfinder Frontier AI Services
- SentinelOne: Wayfinder Frontier AI Services
- Anthropic: Introducing Claude Fable 5.1 and Claude Mythos 5.1
- Anthropic: Expanding the Cyber Verification Program
- Investing.com: SentinelOne adds Anthropic's Claude to security service
