Anthropic is reportedly launching the public version of its Claude 5 AI model, named “Claude Fable,” today, June 9. This new model is designed to be safety-restricted and is expected to cost about twice as much as its predecessor, Claude Opus, but significantly less than the original Mythos pricing. Claude Fable aims to improve the handling of long, complex, multi-turn tasks while incorporating more substantial guardrails and weaker cyber capabilities than those available to a limited group of approximately 200 vetted organizations, which had early access due to safety concerns. Anthropic has not yet confirmed the launch.

Anthropic: Anthropic is an AI research company known for building advanced large language models in the Claude family with a strong emphasis on safety and responsible development. It is reportedly launching Claude Fable as the public version of its Mythos model, featuring improved performance on complex multi-turn tasks alongside heavier safety restrictions compared to internal partner versions.
Alex Heath: Alex Heath is a technology journalist and author of the Sources newsletter focused on the AI industry. He serves as a primary source for reports on Anthropic’s model release plans, highlighting differences in capabilities and access between public and partner-only versions.
Claude Fable: Claude Fable is the name for the upcoming public release of Anthropic’s Mythos model. It is positioned as a safety-restricted version optimized for handling long, complex tasks while limiting certain advanced capabilities available to select partners.

Safety Focus: Anthropic prioritizes guardrails in public model releases to address concerns around advanced capabilities like cyber applications.
Partner Access: Limited early access to full model versions has been restricted to a small number of vetted organizations for safety reasons.
Model Advancement: The development targets better support for intricate, multi-step AI interactions in a broadly available format.