Anthropic Launches Claude Fable 5 With Safeguards, Deploys Claude Mythos 5 With US Project Glasswing
Image: WIRED

Anthropic Launches Claude Fable 5 With Safeguards, Deploys Claude Mythos 5 With US Project Glasswing

09 June, 2026.Technology and Science.19 sources

Key Takeaways

  • Claude Fable 5 is a public Mythos-class model with safeguards for general use.
  • Claude Mythos 5 is available to vetted partners through Project Glasswing for specialized domains.
  • Fable 5 enables long-running agents and multi-stage tasks with guardrails blocking high-risk outputs.

Mythos-class models, gated safeguards

Anthropic on Tuesday launched Claude Fable 5, a “Mythos-class” model made “safe for general use,” and it also released Claude Mythos 5 for a “small group of cyberdefenders and infrastructure providers.”

Key takeaways - Claude Fable 5 makes Mythos-level capabilities available to all customers, with strong safeguards designed to make it safe for broader use

About AmazonAbout Amazon

Anthropic said Fable 5’s safeguards mean that “queries on some topics will instead receive a response from our next-most-capable model, Claude Opus 4.8,” and it added that the safeguards “trigger, on average, in less than 5% of sessions.”

Image from About Amazon
About AmazonAbout Amazon

The company said it is “launch[ing] the model with safeguards” because “Without safeguards, Fable 5’s capabilities in areas like cybersecurity could be misused to cause serious damage.”

Anthropic also said Mythos 5 is “the same underlying model as Fable 5, but with the safeguards lifted in some areas,” and that Mythos 5 will initially be deployed through Project Glasswing in collaboration with the US government.

Fallback to Opus 4.8

Multiple outlets described how Fable 5 routes high-risk requests to Claude Opus 4.8, with The Verge saying Anthropic released Fable 5 “with the system falling back to Claude Opus 4.8.”

The Verge reported that “95 percent of Fable sessions ran entirely on Fable responses, without falling back to Opus 4.8,” while Anthropic’s own launch materials said the safeguards “trigger, on average, in less than 5% of sessions.”

Image from Anthropic
AnthropicAnthropic

Anthropic also framed the approach as conservative tuning, writing that it “tuned these safeguards conservatively—they’ll sometimes catch harmless requests,” and Fortune quoted Dianne Na Penn saying, “The reason why we’re releasing Fable 5 now is very much due to us feeling more confident with our safety guardrails in place.”

Fortune added that responses in specific high-risk areas, “such as biology and cybersecurity,” will be blocked and “in most cases” answered by Claude Opus 4.8, and it said Fable 5 is priced at $10 per million input tokens and $50 per million output tokens.

Access, pricing, and what it can do

Anthropic said Fable 5 is available through Amazon Bedrock, where customers can “build within your existing AWS environment” and access the model via the “Bedrock Runtime console or the new Bedrock Mantle console.”

In its own launch materials, Anthropic said Fable 5 can work “autonomously for longer than any previous Claude models,” and it described early testing where Stripe reported that Fable 5 “compressed months of engineering into days.”

Anthropic also tied the release to long-running work and vision, saying Fable 5 can “understand diagrams, charts, and tables nested in files and PDFs,” and it said the model can plan, “check progress against the goal, and refine its work as it goes.”

On pricing and rollout, Anthropic said both Fable 5 and Mythos 5 are offered at $10 per million input tokens and $50 per million output tokens, and Fortune reported that Fable 5 is rolling out as a general release while Mythos 5 is “to vetted partners” with access to Claude Mythos Preview.

More on Technology and Science