GitHub Copilot usage metrics API now shows usage by tool, including OpenCode and Claude Code

What changed in the Copilot usage metrics API

GitHub's changelog dated October 23, 2026 says the Copilot usage metrics API now reports model usage from more of the tools developers use with Copilot. This includes tools like OpenCode that the reports did not capture before. A new totals_by_client array appears in the enterprise, organization, enterprise-user and organization-user reports, in both the 1-day and 28-day versions. In the aggregated 28-day reports, the array sits inside each day_totals entry.

GitHub says the gap existed because tools outside the IDE often do not send editor telemetry, so their usage was missing from totals_by_ide. The new figures are counted on the server.

Sources: Usage metrics API adds usage from tools like OpenCode - GitHub Changelog

What each client entry reports

Each entry covers one client, meaning the tool that sent the requests. The listed values are opencode, claude_code, codex, zed, copilot_cli, github_app, cloud_agent, code_review, vscode and jetbrains. Clients the API does not recognize are grouped under other.

For each client the report gives request_count (successful model requests), session_count (distinct sessions) and token_usage (prompt tokens, output tokens and average tokens per request). A nested totals_by_model breaks the same figures down by model. Its model_selection_mode field shows user_selected, auto or product_selected. Modes that are not yet reported individually appear as other, and values that cannot be normalized appear as unknown.

Sources: Usage metrics API adds usage from tools like OpenCode - GitHub Changelog

Who can see the new data

Access is limited to enterprise owners and billing managers, organization owners, and anyone with a custom organization or enterprise role that grants the View Copilot Metrics permission. The Copilot usage metrics policy must be enabled. The changelog lists only enterprise and organization reports, at the account level and per user.

Sources: Usage metrics API adds usage from tools like OpenCode - GitHub Changelog

How to read the numbers against usage and spend

request_count counts model calls, not prompts. One prompt can trigger several model calls, so GitHub warns against comparing this field directly with prompt or interaction counts. Each client is identified by the integration ID it sends with every request, and GitHub calls the client field a best-effort grouping, not a verified identity.

Requests from Copilot code review and Copilot cloud agent report product_selected because the product picks the model, and Copilot code review models show as unknown. The changelog describes request and token counts only. It does not say whether the new fields include AI Credits or dollar amounts.

Sources: Usage metrics API adds usage from tools like OpenCode - GitHub Changelog

What stays the same

The change is additive. Existing fields are unchanged, and totals_by_client is an empty array when there is no usage to report. It is separate from agent app activity, which counts jobs started by agent apps such as Claude and Codex on GitHub. totals_by_client measures model requests and tokens from any tool that uses Copilot.

Sources: Usage metrics API adds usage from tools like OpenCode - GitHub Changelog