Anthropic releases Claude Fable 5.1 and restricts Mythos 5.1 access

Anthropic released Claude Fable 5.1 on September 1, 2026, across its platforms and major cloud services. It paired the model with Claude Mythos 5.1, a related variant available only to vetted cybersecurity and life-sciences professionals.

Anthropic cut Fable 5.1’s cache-read price to $0.25 per million tokens, 75% below the previous cache-read price, while keeping other rates at $10 per million input tokens and $50 per million output tokens. Cache reads cover parts of a prompt that a service has already processed and stored, allowing repeated context to be billed at the lower rate instead of the full input price.

Fable 5.1 became available through Anthropic’s API and on Amazon Web Services, Google Cloud and Microsoft Azure. Its Claude API model identifier is claude-fable-5-1. The launch follows the earlier Claude Fable 5 and Mythos 5 release cycle.

Anthropic describes Fable 5.1 and Mythos 5.1 as the same underlying model deployed with different safeguard levels. The company estimates that cheaper cache reads reduce typical usage-based workload costs by about 25% compared with Fable 5 and by about 45% for complex coding and highly agentic workloads. Anthropic said those estimates are based on August 2026 usage.

Access to Mythos 5.1 runs through Anthropic’s trusted-access programs for vetted cyberdefenders and life-sciences professionals and was initially limited to a set of US organizations. Anthropic said its Cyber Verification Program would add Mythos-class access in the near future and that its Life Sciences Verification Program had enrolled initial participants in partnership with the US government. The announcement does not provide a date for either expansion.

The company said Mythos 5.1 applies more permissive cybersecurity and life-sciences safeguards for vetted users while retaining its other safeguards. Fable 5.1 permits defensive software-vulnerability identification but continues to redirect penetration testing, exploit generation and binary-based vulnerability scanning to Opus models.

Anthropic reported that Fable 5.1’s updated cybersecurity safeguards produce about 60% fewer interventions per Claude Code session than the safeguards introduced with Fable 5. That comparison is based on the company’s measurements.

Anthropic said its evaluations placed Mythos 5.1 below the next chemical-and-biological risk tier under its Responsible Scaling Policy and within the lower cyber-risk category of its Frontier Compliance Framework. The release material does not include an independent reproduction of those evaluations.

Anthropic also said its Enterprise Frontier Safeguards, which store customer data in customer-controlled cloud infrastructure, will begin a phased rollout in fall 2026. Eligible customers can use Fable 5.1 with zero data retention until that rollout reaches them.