Anthropic cuts Claude Fable 5.1 costs, adds enterprise data controls
Serge Bulaev
Anthropic released Claude Fable 5.1 and Mythos 5.1, which may lower costs for users and offer two security levels. Fable is open to the public, while Mythos is only for selected users. The company says typical bills might drop by 25 percent due to lower cache read costs, and Fable 5.1 appears to perform much better on science tests than earlier versions. Early feedback suggests people like the model's quality, but some worry about how quickly usage limits are reached. New data controls let customers keep their usage data in their own cloud storage, which may help with privacy and compliance needs.

In a key move, Anthropic cuts Claude Fable 5.1 costs and adds new enterprise data controls with the release of two new models. The update introduces Claude Fable 5.1 for public use and the restricted Mythos 5.1, focusing on performance, cost savings, and enhanced privacy for businesses.
This report reviews the pricing change, benchmark gains, customer reactions, and new enterprise privacy controls.
Cost and performance shifts after Anthropic releases Claude Fable 5.1 and Mythos 5.1 with tiered safeguards
Anthropic's Claude Fable 5.1 update significantly reduces costs by cutting cache read prices by 75%, potentially lowering typical bills by 25%. It also doubles performance on key science benchmarks and introduces Enterprise Frontier Safeguards, giving businesses greater control over their data privacy and storage.
A major cost-saving measure is the 75% reduction in cache read pricing, now at $0.25 per million tokens. Anthropic estimates this change can lower typical bills by 25% and up to 45% for intensive agentic workflows, detailed in the Anthropic launch post, citing improved cache hit rates.
In performance, Fable 5.1 shows dramatic improvement, scoring 52.6% on the Terminal-Bench-Science 0.1 benchmark - more than double the 24.7% of its predecessor, Fable 5. This score significantly surpasses competitors like Opus 5 (29.0%) and GPT-5.6 Sol (22.4%), a finding confirmed by an independent RD World Online analysis.
The model's safeguards have also been refined. Cyber-related safety triggers have been reduced by approximately 60%, minimizing false refusals for technical content. Internal testing also indicates superior code generation and cleaner diagram reproduction.
Early user feedback and adoption signals
Initial public feedback highlights a mix of enthusiasm and concern. While users praise Fable 5.1's capabilities, a widely shared Reddit discussion thread points out that its high performance can rapidly consume usage quotas. Reviewers have noted the model's "almost universally positive" reception for writing and reasoning but echo worries about pricing and access.
According to industry reports, internal evaluations show significant preference for Fable 5.1 over its predecessor on standard benchmarks. In early adoption signals, enterprise clients have begun migrating workloads to the new model. While official adoption metrics are not yet available, analysts anticipate strong uptake within Anthropic's user base.
Enterprise Frontier Safeguards and data control
A central feature for businesses is the new Enterprise Frontier Safeguards (EFS). This allows customers to store all model usage data within their own cloud environments, such as Amazon S3 or Azure Blob Storage. Anthropic positions this as offering privacy on par with zero data retention, while still enabling automated misuse detection. According to reports, customers will control data review, storage, and encryption at no extra cost.
Key details for security and compliance teams:
- Storage location - customer cloud rather than Anthropic infrastructure
- Default human review - performed by the customer, not Anthropic
- Retention window - reportedly remains 30 days for some tiers, but custody shifts to the client
- Rollout - phased availability beginning later this fall
Importantly, reports suggest the 30-day data retention window may still apply for certain models; the primary change is data custody, not retention duration. This shift is expected to simplify procurement for organizations in regulated sectors like finance and healthcare, which often have strict policies against vendor-side data storage.
What happens next
Looking ahead, Anthropic will roll out Enterprise Frontier Safeguards in phases while continuing to gather user feedback on pricing, usage limits, and model access. Fable 5.1 and Mythos 5.1 will serve as the company's flagship models as it explores further refinements to its safeguards and performance benchmarks.