Anthropic Releases Claude Fable 5.1 With Steep Price Cuts for Agent Workflows

Anthropic rolled out Claude Fable 5.1 and Mythos 5.1, updating its model lineup with price cuts and recalibrated guardrails designed to appease enterprise developers building autonomous agents. The release directly targets recurring complaints around operating expenses, data privacy, and overzealous safety filters, according to a report by The Verge.

According to Anthropic, Fable 5.1 delivers stronger overall performance than Fable 5 while lowering standard usage costs by roughly 25 percent. For complex agentic workflows that rely heavily on processed and stored prompt caching, expenses drop by as much as 45 percent, easing the operational overhead of running persistent tools.

Early enterprise testers reported notable gains in tone and nuance. Dan Shipper, chief executive of Every, described Fable 5.1 as the strongest coding system his team has evaluated, emphasizing that it is fast, token-efficient, and natural in its conversational cadence. Aaron Levie, chief executive of Box, said his team's agent successfully caught ambiguities and data subtleties that the previous iteration overlooked in identical tests.

Anthropic also adjusted its safety systems to reduce false positives. Fable 5.1 features more granular boundaries that avoid improperly blocking benign biology queries, though Mythos 5.1 retains earlier restrictions in that domain. Furthermore, Anthropic is now permitting Fable 5.1 to scan for software vulnerabilities, though high-risk security workloads such as binary scanning, exploit creation, and automated penetration testing will still route to Opus models.

To address corporate data privacy concerns, the startup outlined Enterprise Frontier Safeguards, an architecture scheduled to launch later this fall that retains customer data entirely on client cloud infrastructure rather than Anthropic systems.

The updates arrive as frontier AI labs increasingly compete on developer unit economics rather than raw parameter count alone. By addressing overactive refusals and lowering token caching fees, Anthropic is attempting to defend its market share among software engineering teams whose multi-step agent loops become cost-prohibitive under standard API pricing tiers.

Claude Fable 5.1 is now generally available across all standard Anthropic interfaces and APIs, while Mythos 5.1 deployment remains restricted to participants in the company's Project Glasswing initiative.