Anthropic Introduces Model Access Control for Enterprise Administrators

Anthropic published a new feature for organizations on the Enterprise plan in its customer support center on July 1, 2026: model entitlements in beta. This is an administrative control that gives enterprise admins precise management over which Claude models are available to users within the organization, as well as the ability to configure effort-level settings.

What Does the New Feature Bring?

Until now, all users within an enterprise organization had equal, undifferentiated access to the Claude models available under the plan. The new beta feature changes this on two levels.

First, administrators can now restrict or define which specific models users have access to. In practice this means, for example, that the customer support team can be limited to faster and more economical models, while engineers or analysts retain access to more advanced versions. Organizations with specific compliance requirements gain a tool to ensure that certain models are not used on sensitive data streams.

Second, admins can manage effort-level settings — the parameter that determines how much computational resources Claude invests in a given task. A higher effort level delivers more thorough but slower responses and higher costs per query. The ability to set these values per user or group gives organizations finer control over the cost structure and response speed for different teams.

Context: A Follow-Up to Rate-Limit Tier Consolidation

This beta is not an isolated step. A week earlier, on June 26, 2026, Anthropic announced the consolidation of rate-limit tiers under new names: Start, Build, and Scale. That change standardized the usage-limit structure at the plan level. Model entitlements are a logical continuation of that reform: while the new limit structure defines how much can be used, entitlements define what may be used and how.

Together, these two changes build a more consistent framework for enterprise governance of Claude — something that larger organizations, especially those in regulated industries, have long sought from AI service providers.

Who Manages Model Access

According to available documentation, the controls are reserved for administrators on Enterprise plans. Entitlements can be set at the level of an individual user or a group of users within the organization. Anthropic has published a dedicated guide titled Manage model access for your organization in the customer support center, which explains the setup to administrators in detail.

Beta Limitations

It is important to note that the feature is still in beta. This means the interface and capabilities may change, and temporary unavailability or behavior changes are possible. Organizations that plan to rely on this control in production environments should follow Anthropic’s release notes to stay informed of any changes.

Significance for Enterprise Users

The ability to granularly control model access rarely stands out as spectacular news, but for IT and security teams in organizations it represents practical value. It reduces the risk of unauthorized use of more advanced and more expensive models, makes cost planning easier, and enables organizations to support their own internal AI tool usage policies through platform enforcement.

For now, this is a modest but useful addition to the administrative toolkit on the Claude Enterprise plan.