Copilot usage metrics API adds BYOK, code review addressed rate, and telemetry gap fields

GitHub shipped three additions to the Copilot usage metrics API between November 3 and 5. They help enterprises see bring your own key (BYOK) usage, measure whether Copilot code review feedback is acted on, and find users whose reports are incomplete.

BYOK and enterprise custom models (November 3)

• Local BYOK: Developer-configured keys and local models (such as Ollama, Foundry Local, LM Studio, and llama-server) used from Copilot CLI or the Copilot app. New fields include used_local_byok, totals_by_local_byok (broken out by byok_kind and surface), and daily, weekly, and monthly active local BYOK user counts.
• Enterprise custom models: Admin-configured models routed through GitHub. A new boolean is_custom_enterprise_model appears on totals_by_client[].totals_by_model[] entries and covers every client, including IDEs.
• Scope: Enterprise and organization reports, user-level and aggregate, 1-day and 28-day. Existing totals do not change and there is no backfill.
• Caveats: Local BYOK relies on client telemetry, so it is a lower bound, and some local providers do not report tokens. Custom model names that are not recognized appear as unknown.

Code review details and addressed rate (November 5)

• A new copilot_code_review object on each repos-1-day row of the repository-level reports.
• total_comments_evaluated and total_comments_addressed give an addressed rate per repository. A comment counts as addressed if the developer applied the suggestion, asked Copilot to fix it, edited the code, or replied that the fix will come later.
• Breakdowns by severity, review tier (lite, balanced, max), comment type (such as bug, security, custom), and customization (custom instructions, skills, MCP servers, or none).
• total_merged_no_human_change_request_after_review counts merged Copilot-reviewed pull requests with no later change request from a person.
• Comments are evaluated when the pull request merges, and comments on bot-opened pull requests are usually not evaluated.

Telemetry and version gaps (November 5)

• Per-user client_telemetry_present and usage_metrics_requirements (each IDE or extension marked met, unmet, or unknown against minimum versions).
• Daily counts missing_client_telemetry_user_count and below_usage_metrics_minimum_user_count. These overlap, so they should not be added together.
• Users without telemetry still count as active but lack details such as lines of code and model breakdowns. Common causes are old editors, disabled IDE telemetry, and proxies or firewalls.

Access

Enterprise owners, enterprise billing managers, organization owners, and holders of a custom role with the View Copilot Metrics permission, with the Copilot usage metrics policy enabled.