Anthropic has released Claude Fable 5.1 and Claude Mythos 5.1, describing them as its most advanced models for coding and knowledge work. The company says the two products are the same underlying model shipped with different safeguard configurations: Fable 5.1 is generally available, while Mythos 5.1 stays limited to trusted access programs aimed at cybersecurity and life-sciences research.
On the commercial side, Anthropic is cutting the price of cache reads by 75% wherever usage is billed by token. The company estimates that change lowers typical workload cost by about 25% versus Fable 5, and by up to about 45% for highly agentic, context-heavy jobs. Base list pricing otherwise remains $10 per million input tokens and $50 per million output tokens. Developers can call Fable 5.1 as claude-fable-5-1 on the Claude API, and it is also available through Amazon Web Services, Google Cloud, and Microsoft Azure.
Anthropic's published evaluations put Fable 5.1 ahead of Fable 5 across several agentic and knowledge-work benchmarks. On Terminal-Bench-Science 0.1, Anthropic reports 52.6% for Fable 5.1 versus 24.7% for Fable 5 under its setup. Terminal-Bench 4.0 scores rise to 55.8% for Fable 5.1 and 60.9% for Mythos 5.1, compared with 42.0% for Fable 5. The company also highlights stronger results on knowledge-work, computer-use, and business-workflow suites, while noting that production safeguards can zero some tasks and therefore understate raw capability.
Beyond coding, Anthropic showcased early scientific research demos with the new models. Examples include Mythos 5.1 designing high-affinity protein binders that external labs validated, Fable 5.1 producing a higher-resolution elevation map of part of Venus from Magellan radar data, and custom GPU-kernel work that sped up several open-source biology models by as much as 2.5×. Those claims remain company-reported case studies, but they frame the release as much about research assistance as chat performance.
Safety positioning is central to the launch. Mythos 5.1 keeps more permissive cyber and biology settings for vetted U.S. organizations through Anthropic's Cyber Verification Program and Life Sciences Verification Program. Fable 5.1 keeps tighter defaults, though Anthropic says cyber false-positive interventions fall by about 60% and that Fable 5.1 may now be used to identify software vulnerabilities without generating exploits. Enterprise Frontier Safeguards, rolling out in phases later this fall, are meant to give eligible customers zero-data-retention-style privacy by storing reviewable data in customer-controlled cloud infrastructure.
Availability is already live for Fable 5.1 on Anthropic's platforms and major clouds. Mythos 5.1 remains invitation-only. Independent coverage from AWS and model-evaluation outlets corroborates the general-availability split, the API model ID, and the framing of Fable and Mythos as safeguard variants of one underlying system rather than two separate trainings.