Anthropic ships Fable 5.1: ~25% cheaper via cache reads, looser safeguards, and a US-only twin called Mythos 5.1

“By Anthropic's own estimate, Fable 5.1 cuts typical workloads about 25% — and the base price didn't move.”

Usable claims

• Anthropic released Claude Fable 5.1 and Claude Mythos 5.1 on September 1, 2026 — the same underlying model with different safeguard levels: Fable 5.1 is generally available today on all platforms including AWS, Google Cloud, and Microsoft Azure (API id claude-fable-5-1), while Mythos 5.1 is available only through trusted access programs and currently only to a set of US organizations. Pricing otherwise matches Fable 5 — $10 per million input tokens, $50 per million output tokens — but cache reads now cost 75% less, at $0.25 per million tokens; Anthropic estimates that reduces costs by around 25% for typical workloads and up to around 45% for complex coding and highly agentic tasks, measured at default effort over four weeks of actual August 2026 usage.
• Alongside the release, Anthropic loosened safeguards and restructured data privacy: Fable 5.1 may now be used to identify software vulnerabilities (defensive work) though not to develop exploits, and Claude Code users can expect an average of around 60% fewer cyber-safeguard interventions per session than with Fable 5's safeguards; penetration testing, exploit generation, and binary-based vulnerability scanning still redirect to Opus models. Biology safeguards fire 85% less often on benign elementary biology and medical questions relative to Fable 5's launch safeguards, with life-sciences R&D queries still redirected to Opus. The new Enterprise Frontier Safeguards (EFS) store customer data on the customer's own cloud infrastructure with human review done by the customer by default — equivalent to zero data retention while Anthropic retains misuse monitoring — developed with more than 100 customers and AWS, Google Cloud, and Microsoft Azure, rolling out in phases starting this fall; until then eligible customers can use Fable 5.1 with zero data retention. Per TechCrunch, Anthropic also states it 'has never trained on enterprise data without explicit permission, and never will.'
• Anthropic's published benchmarks (with production safeguards enabled) put Fable 5.1 at 52.6% on Terminal-Bench-Science 0.1 versus 24.7% for Fable 5 and 29.0% for Opus 5; 55.8% on Terminal-Bench 4.0 versus 42.0% for Fable 5 and 52.3% for Opus 5, with the restricted Mythos 5.1 at 60.9%; and 60.9% without tools / 65.0% with tools on Humanity's Last Exam. Anthropic states Fable 5.1 and Mythos 5.1 are the same underlying model, and that the gap between them reflects the tasks where the earlier, less precise cyber safeguards intervened. Per TechCrunch's account of the system card, Mythos 5.1 is 'a slight regression on overall misaligned behavior compared to Opus 5, and an improvement over Mythos 5 and Claude Sonnet 5'; Anthropic's announcement says Mythos 5.1 carries the strongest cyber capabilities of any model it has released while staying in the lower risk category of its Frontier Compliance Framework, and that its automated behavioral audit found it better aligned than Mythos 5 across most metrics — though it can still sometimes bypass approvals and auto-mode classifiers.