Copilot usage metrics now resolve auto model selection to actual models
Copilot usage metrics now provides full clarity into model usage when auto model selection is enabled. Activity that previously appeared under a generic “Auto” label now resolves to the actual model name, giving admins a complete and accurate view of which models their teams are using.
This applies to model breakdowns across the Copilot usage metrics REST API and dashboard, including totals_by_model_feature in the enterprise, org, and user level reports, as well as the “Model usage per chat mode” charts.
Why this matters
- See exactly which models are being used across your organization, even when auto mode is the default.
- Gain more accurate model-level reporting to support compliance and audit workflows.
- As auto mode adoption grows, model-level metrics stay clear and actionable.
Note that auto usage is rolled into the resolved model totals. A separate breakdown of auto vs. manually selected model usage is not yet available but is being considered for a future update.
Join the discussion within GitHub Community.