Anthropic releases Claude Fable 5, a gated version of Mythos for public access
Anthropic brings its most powerful model to the general public through Claude Fable 5, paired with safety guardrails and mandatory 30-day traffic retention.
Last verified:
Claude Fable 5 debuts with hard safety guardrails
Anthropic released Claude Fable 5 on June 10, making its Mythos-class model available to the general public for the first time, according to TechCrunch. The launch marks a shift from Mythos’s initial April preview, when access was restricted to a small number of approved partners and later expanded to hundreds of organizations across 15 countries managing critical infrastructure.
Fable 5 excels at software engineering, knowledge work, and vision tasks, but delegates requests in high-risk domains to Claude Opus 4.8. According to TechCrunch, these restricted areas include cybersecurity, biology, chemistry, and model distillation. Early data indicates at least 95% of user sessions run entirely on Fable 5 without deferring to Opus 4.8, meaning safety blocks are infrequent in practice.
Staged rollout and pricing structure
Access follows a two-phase timeline. Through June 22, Fable 5 is included at no additional cost in Claude Pro, Max, Team, and seat-based Enterprise plans. On June 23, Anthropic will transition the model to a consumption-based credit system, requiring users to pay per use. The company plans to restore Fable 5 as a standard subscription feature as soon as possible, pending further evaluation.
Concurrently, Anthropic is deploying Mythos 5 to organizations already approved for advanced model access, though the source does not specify whether Mythos 5 carries the same safety restrictions as Fable 5.
Safety testing and data retention
Before release, Anthropic stress-tested its classifiers with jailbreak attempts, according to TechCrunch. The company also engaged external red-teaming organizations, neither of which identified universal jailbreaks across over 1,000 hours of testing.
To mitigate emerging attack vectors, Anthropic is enforcing a mandatory 30-day traffic retention policy on all Fable 5 and Mythos 5 usage, even for enterprises with prior zero-retention agreements. According to TechCrunch, the company states this data will not be used for training but only for defending against novel attacks and reducing false positives in safety classifiers.
Why This Matters
The public release of a Mythos-class model signals Anthropic’s confidence in its safety architecture, but the mandatory traffic retention requirement may set a precedent: access to frontier models increasingly comes with conditions that prioritize security over user privacy. Organizations deploying Fable 5 should evaluate whether the 30-day retention aligns with their data governance policies. The staggered pricing transition also suggests Anthropic intends to meter access to Mythos-class capability, treating it as a premium offering rather than a commodity feature—a stance that may influence how competitors price frontier models.
Frequently Asked Questions
What is Claude Fable 5?
Fable 5 is the first publicly released version of Anthropic's Mythos model, available through the Claude API and Enterprise plans. It excels at software engineering and knowledge work but defers sensitive requests to Claude Opus 4.8.
How does safety gating work in Fable 5?
The model blocks responses in high-risk areas including cybersecurity, biology, chemistry, and model distillation, automatically falling back to the less-capable Opus 4.8 for those requests.
Why is Anthropic requiring 30-day traffic retention?
Anthropic says the policy helps defend against novel jailbreak attempts and reduce false positives in safety classifiers—a measure the company frames as essential for increasingly powerful models.