GitHub adds per-user AI credit usage to Copilot metrics API

GitHub has added per-user AI credit consumption data to its Copilot usage metrics API, introducing an ai_credits_used field in user-level reports for enterprises and organizations.

What changed

In a June 19 GitHub Changelog post, GitHub said the new ai_credits_used field is now available in Copilot usage metrics API user-level reports. The field is intended to help enterprise administrators and organization owners connect Copilot usage with adoption, value measurement, and budget planning.

The update applies to user-level reporting rather than only aggregate usage views. According to GitHub’s changelog, that means administrators can see how many AI credits were consumed by individual Copilot users in the relevant metrics reports.

GitHub’s REST API documentation for Copilot usage metrics lists endpoints for retrieving Copilot metrics at the enterprise, organization, team, and user levels. The documentation includes enterprise user-level reports for a specific day and reports covering the latest 28 days of user usage metrics. These API endpoints are designed for administrators who need programmatic access to Copilot usage data rather than relying only on dashboards.

Why it matters for administrators

The addition of ai_credits_used gives administrators a more direct way to compare Copilot activity with consumption-based billing data. GitHub’s changelog frames the change around adoption, value, and budget planning, which reflects a common challenge for companies deploying AI coding tools: tracking whether usage is spreading across teams, and how that usage maps to costs.

GitHub’s billing documentation separately describes billing usage endpoints, including AI credit usage reports. The billing documentation shows Copilot AI Credits returned with credit units, supporting the link between Copilot usage reporting and credit-based billing analysis.

Together, the Copilot metrics API and billing usage API can help enterprises build internal reporting that connects who is using Copilot, how often they are using it, and how AI credit consumption is trending. GitHub has not said in the changelog that the new field changes pricing or entitlement rules; the announced change is the addition of a reporting field.

Where the field appears

GitHub’s changelog specifically describes the new field as part of Copilot usage metrics API user-level reports. GitHub Docs for the Copilot usage metrics REST API describe endpoints for user-level reporting, including enterprise reports for a specific day and latest 28-day reports.

That distinction is important for administrators building reporting systems. Aggregate Copilot adoption numbers can show general usage patterns, but per-user credit consumption can support more detailed questions, such as whether usage is concentrated among a small group, whether a rollout is reaching intended teams, or whether credit consumption is aligned with expected business needs.

Broader context

GitHub Copilot has expanded from code-completion features into a broader set of AI-assisted developer tools, and GitHub’s documentation now separates product usage metrics from billing usage reporting. The new ai_credits_used field narrows the gap between those two reporting areas by making consumption visible at the user level inside Copilot metrics reports.

For enterprises, this kind of reporting can be useful for chargeback models, internal forecasting, and governance reviews. However, the sources describe the feature as a metrics addition, not as a recommendation for how organizations should evaluate developer productivity or individual performance.

The practical takeaway is straightforward: GitHub enterprise administrators and organization owners who already use the Copilot usage metrics API can now include per-user AI credit consumption in their reporting, while GitHub’s billing usage endpoints remain the source for broader AI credit billing data.