Claude Fable 5

Model Overview

Claude Fable 5 is Anthropic's most capable publicly available frontier model as of mid-2026. Released on June 9, 2026, it is engineered for long-running, multi-agent autonomous workflows and sustained logical reasoning. It belongs to the "Mythos-class" of models and represents a significant step up from the Claude Opus series.

Capabilities

• Long-Horizon Reasoning: Capable of maintaining context and logic over extended, complex tasks (e.g., multi-day coding sessions).
• Agentic Workflows: Native multi-agent team coordination for advanced knowledge work and software engineering.
• Multimodal Interpretation: Advanced vision capabilities for interpreting charts, diagrams, and files alongside text and code.
• Large Context Window: 1 million token context window.

Example Use Cases

• Complex software engineering and autonomous coding tasks.
• In-depth document analysis over large knowledge bases.
• Autonomous, long-running agentic tasks.
• Advanced logic and problem-solving scenarios.

Performance & Benchmarks

Fable 5 shares the underlying architecture with the highly restricted Claude Mythos 5, bringing "Mythos-level" capabilities to the general public. While specific quantitative benchmark numbers are typically undisclosed, it significantly exceeds the performance of earlier models like Claude Opus 4.8. Built-in safety classifiers may cause it to decline requests or fall back to Claude Opus 4.8 for queries touching on high-risk topics like cybersecurity, biology, or chemistry.

Intended Use & Limitations

Designed for Pro, Team, and Enterprise users handling complex workflows. The model includes robust safety safeguards, meaning it will refuse to assist with or participate in activities that trigger its security classifiers. It is priced at $10 per million input tokens and $50 per million output tokens.

About Anthropic

Anthropic is an AI safety and research company focused on building reliable, interpretable, and steerable AI systems. Their flagship product is the Claude family of AI models, emphasizing safety and alignment.