Anthropic has released Claude Fable 5, the company’s first public model to bring Mythos-class capabilities to enterprise customers. Announced on June 9, 2026, Fable 5 represents a significant expansion of Anthropic’s enterprise offerings, which had previously been limited to preview access for select partners.
The release includes three major enterprise features. First, Managed Agent Sandboxes provide secure, isolated environments for running autonomous agents without exposing sensitive corporate data. Second, 20+ Legal MCP Connectors enable enterprises to integrate Claude with existing legal workflows, contract management systems, and compliance tools. Third, Apple Foundation Models Integration brings Claude’s capabilities to Apple enterprise ecosystems, with deep integration into iOS and macOS device management.
“We’re seeing enterprises move beyond experimentation to production deployment,” said an Anthropic spokesperson. “Fable 5 is designed for that transition—from proof-of-concept to real business value.”
The model retains the strong computer-use and browser-agent capabilities that defined Opus 4.8, while adding enterprise-specific features including audit logging, role-based access controls, and SOC 2 compliance readiness. Early enterprise adopters include financial services firms and healthcare organizations that require strict data governance.
Fable 5’s release coincides with the June 15 retirement of Opus 4 and Sonnet 4, signaling Anthropic’s commitment to a streamlined model family. The company has positioned Fable 5 as the primary enterprise choice, with Sonnet 5 serving the mid-tier market and Haiku handling high-volume, low-latency workloads.
The timing of Fable 5’s release appears strategic, competing directly with OpenAI’s GPT-5.6 and Google’s Gemini 3.2 in the enterprise market. Industry analysts note that Anthropic’s focus on enterprise-specific features—particularly managed sandboxes and legal MCP connectors—differentiates its offering from rivals targeting the broader developer market.