Microsoft provides a new admin setting to control AI models in preview. Preview and experimental models are intended for exploration and testing and are not recommended for production use.
The new setting currently supports Anthropic models with data retention, including the Claude Fabel 5 model for Copilot Cowork (Frontier). Fabel 5 is a Mythos-class model, a model with data retention, as described by Anthropic.
To ensure we’re responsibly deploying Mythos-class models, we are requiring limited data retention and review as part of our safety work. Prompts submitted to, and outputs generated by, Mythos-class models are retained for 30 days for trust and safety purposes, on every platform where these models are offered.
This applies to Mythos-class models and future models with similar capabilities that we designate as covered models. For all other models, everything you use is unaffected and stays under the current terms.
Mythos-class models can be controlled in the Microsoft 365 admin center under Copilot > Settings > AI models in preview. The Anthropic setting is disabled by default.
