Claude Fable 5.1: Anthropic's New Frontier Model for Coding and Knowledge Work

Anthropic has released Claude Fable 5.1 and Claude Mythos 5.1, marking another significant step forward in the model race. These are the company's most advanced models yet for coding, knowledge work, and long-running problem-solving tasks.

What Changed

Claude Fable 5.1 is a strict performance improvement over its predecessor, Fable 5. The key upgrades include:

• Stronger coding performance: Set a new standard on Terminal-Bench 4.0, the agentic search benchmark, and long-running problem-solving tasks
• Improved cost-efficiency: At Low or Medium effort levels, Fable 5.1 achieves results similar to or better than Fable 5's high-effort mode at a fraction of the cost
• Enhanced cyber capabilities: Demonstrates the strongest cyber capabilities of any model Anthropic has released, while still staying within the lower category of risk in their Frontier Compliance Framework
• Better distillation resistance: Strengthened mechanisms make it harder for attackers to distill capabilities from the model

Mythos 5.1: The Trusted Access Version

Claude Mythos 5.1 is the same underlying model as Fable 5.1, but with different safeguards tailored for high-risk work:

• Cybersecurity: Specifically designed to support work in cybersecurity while preventing dangerous capabilities
• Life sciences: Safeguards that restrict access to research biology capabilities, following the same approach as Mythos 5
• Trusted access only: Available exclusively through Anthropic's trusted access programs for organizations with specific needs

Pricing and Availability

Claude Fable 5.1 is available today across all Anthropic platforms, including:

• Amazon Web Services
• Google Cloud
• Microsoft Azure
• The Claude API

• Input: $2 per million tokens
• Output: $10 per million tokens

Why This Matters for Builders

Fable 5.1's agentic capabilities mean developers can now handle more complex autonomous workflows without needing to escalate to expensive Opus-tier models. The model's improved cost-efficiency at lower effort levels makes it practical for long-running tasks like code reviews, documentation generation, and multi-step research workflows.

The strengthened distillation defenses address a growing concern as frontier models become more accessible to researchers and open-source communities.

What's Not Here

This is a model release announcement, not a major product platform shift. Claude Code, Cowork, and other existing interfaces continue to work as before. The release represents a capability upgrade rather than a new product category or significant pricing disruption for existing customers.

• https://www.anthropic.com/claude-fable-and-mythos-5-1
• https://www.anthropic.com/news