Microsoft Restricts Employee Access to Anthropic’s Claude Fable 5 AI Model

By: Harsh Vardhan

|

Updated on: 11-Jun-2026 01:00 PM

9,693 views

Follow Us:

9,693 views

Microsoft has restricted its employees from using Anthropic’s Claude Fable 5 AI model on the company’s internal version of GitHub Copilot. This decision comes despite Microsoft offering the Mythos-class model to external customers through GitHub Copilot and Microsoft Foundry.

Key Highlights

• Microsoft blocks employee access to Anthropic’s Claude Fable 5 on internal GitHub Copilot.
• Data retention policy is main reason for restriction, not cost.
• Other Claude models with zero retention remain available to Microsoft staff.
• Anthropic’s Mythos-class model stores data for 30 days, up to two years if flagged.
• Microsoft legal teams are reviewing Anthropic’s data policy for compliance concerns.

Microsoft Limits Internal Use of Fable 5

According to a report from The Verge, Microsoft employees cannot access Claude Fable 5 internally. The restriction is not due to cost, even though the Mythos-class model costs twice as much as Anthropic’s Opus models. Instead, the main concern is Anthropic’s data retention policy for Mythos-class models.

Anthropic’s policy states that prompts and outputs from Mythos-class models are retained for 30 days. This retention supports safety classifiers, which monitor for novel attacks and jailbreak attempts. After 30 days, Anthropic deletes the data, and it is not used for model training. However, if prompts or outputs are flagged for safety investigations, legal reasons, or policy violations, the data may be kept for up to two years.

This change has raised legal and compliance concerns for enterprise users. Previously, enterprise AI tools often offered zero-retention options, ensuring conversations disappeared after use. Microsoft’s legal teams are now reviewing Anthropic’s policy, focusing on the risks to customer data and confidential information. It is unclear if Claude Fable 5 will become available for broader internal use at Microsoft.

Other Claude Models Remain Available

Other Claude AI models from Anthropic remain accessible to Microsoft employees. These models operate under zero data retention rules, which do not store prompts or outputs after use. This difference in data handling is a key factor in Microsoft’s decision to restrict Fable 5 internally.

Some users online have commented on the restriction, suggesting Microsoft employees may be disappointed by the decision. Jokes and memes circulated, with some users suggesting the company is avoiding high costs, while others focused on the impact for developers who hoped to use Fable 5 for coding tasks.

Anthropic’s Claude Fable 5 and Industry Response

Anthropic recently launched Claude Fable 5, its first Mythos-class model available to the public. The model’s data retention policy marks a shift from previous enterprise AI offerings. Anthropic CEO Dario Amodei has also addressed questions about the model’s use, including its role in sensitive situations.

The introduction of Claude Fable 5 comes as companies like Microsoft reconsider their use of AI tools. Rising costs and questions about productivity gains have led to more cautious approaches. Microsoft’s decision highlights the growing importance of data privacy and compliance in enterprise AI adoption.