Overview
The GitHub Copilot Metrics dashboard shows how GitHub Copilot is adopted and used across engineering teams, what it produces, what it costs, and the value it returns. It covers developer activity, acceptance rates, models and IDEs, the Copilot CLI and agents, pull requests and code review, seat utilization, billing, and ROI. The dashboard is designed for engineering leaders, platform teams, and administrators who need to measure AI coding adoption and evaluate the impact of GitHub Copilot on developer productivity. To open it, in the left navigation panel expand AI Coding Tools and click GitHub Copilot. The dashboard selector at the top left switches between the GitHub Copilot, Claude Code, Cursor, and Windsurf dashboards. To set up the data source, see GitHub Copilot. The dashboard is organized into 13 tabs. Each tab is described below.- Data comes from GitHub’s Copilot APIs and is refreshed on each sync. GitHub publishes each day’s usage the following day, so the dashboard runs one day behind.
- User counts such as daily active users are GitHub’s distinct counts. They are never added up across days; a week or month shows the average day as a whole number.
- Usage appears only for members who have Copilot telemetry enabled in their IDE.
Controls and Filters
What is it?
Interactive controls used to customize dashboard data and reporting views.Where does it exist?
Top section of the dashboard, above the tabs. On small screens they collapse into a Filters button.Features
Project Filter Select a project to view the Copilot metrics of the GitHub organization or enterprise connected to it. The list shows the AI Coding projects that have a GitHub Copilot connection. All (the default) shows every project with a Copilot connection together. Date Range Filter Define a custom reporting period. Future dates cannot be picked. Picking a range clears the quick time filter. Quick Time Filters Predefined reporting windows: 7 Days (the default), 15 Days, 90 Days.- 7 Days and 15 Days cover the last 7 or 15 days, and trend charts show one point per day.
- 90 Days covers the current calendar month and the two before it, and trend charts show one point per month.
Working with charts and tables
- Each chart and table has a ⋮ menu to export just that chart or table to CSV, Excel, PDF, or Image.
- Charts with an ⓘ icon show how the figure is calculated when you hover over it.
- Ranked bar charts show the first 10 bars. View Full List opens every bar on its own page, with a search box and a Back button. Some ranked charts also have a search box that filters their bars.
- Click a table column header to sort by it; click it again to reverse the order. Long tables show 10 rows and scroll.
Executive Overview
What is it?
Headline Copilot figures for the selected project and period: who used it, what it produced, and what the seats cost.Metrics Included
Active Copilot Developers Developers who used Copilot in the period, with the number active on an average day. Code Generations Code suggestions Copilot generated, with how many were accepted. Acceptance Rate Accepted suggestions as a percentage of generated suggestions. Seat Utilization Seats active in GitHub’s current billing cycle as a percentage of assigned seats, for example24 active of 28 seats. Taken from the latest seat count in the period.
Total Cost
Net amount GitHub billed for Copilot in the period.
Cost / User
Total cost divided by the developers who used Copilot.
Acceptances / $1
Accepted suggestions per dollar billed.
Pull Requests
Pull requests created in the period, with how many were merged and how many Copilot reviewed.
AI Credits Used
AI credits consumed in the period, with their net and gross amounts.
Sessions
Copilot client sessions, with the number of requests.
Tokens
Total tokens used, split into prompt and output tokens.
All Developers
Every developer with Copilot activity, with their productivity score, acceptance rate, active days, acceptances and lines added per active day, code generations, and adoption phase. Search by name, login, or email with Search user….
The productivity score is the developer’s percentile in the group: interactions, acceptances, and lines added are added together and multiplied by active days, and developers are ranked on the result. It is shown in points: green from 75, amber from 50, red below 50.
Daily Active Users
GitHub’s distinct Copilot users per day.
Use Cases
- Get a one-screen view of Copilot adoption and value
- Support executive reporting
Adoption & Usage
What is it?
Active users as GitHub counts them, which developers are new, and which Copilot features and surfaces they use.Metrics Included
Daily Active Users Distinct Copilot users per day. New vs Returning Users Developers whose first Copilot day in the data G360 has collected falls in the day (or month), against developers who used it before. Weekly Active Users (Rolling) GitHub’s 7-day active user count on each day. Monthly Active Users by Surface Monthly active users on the last day of the period, for Copilot overall (All Copilot) and for Chat, Agent, VS Code agent, CLI, Copilot app, Cloud agent, and Code review. Feature Adoption % For each Copilot feature, the share of active users who engaged with it on 2 or more days of GitHub’s latest 28-day window. Search features with Search feature…. AI Adoption Phases GitHub’s adoption cohorts on the latest day: for each phase, the users in it over 28 days, how many are engaged, and their average acceptances and lines added per user.Use Cases
- Monitor adoption trends and onboarding of new users
- See which Copilot features and surfaces developers actually use
Productivity
What is it?
What Copilot produced: interactions, code generated and accepted, and lines added and deleted.Metrics Included
Acceptance Rate Accepted ÷ generated suggestions, per day. Code Generations vs Acceptances Suggestions generated and accepted, per day. Interactions User-initiated Chat and agent interactions, per day. Lines Added vs Deleted Lines of code added and deleted with Copilot, per day.Use Cases
- Track AI-assisted coding activity over time
- Evaluate how useful Copilot’s suggestions are
Model Utilization
What is it?
Copilot activity split by model, language, IDE, and feature. Each split is a separate GitHub breakdown and is never added to another.Metrics Included
Model Distribution Share of interactions per model. Model Family Trend Each model family’s share of interactions and code generations, per day. Languages Code generations per programming language. Search with Search language…. IDEs Interactions plus code generations per IDE. Search with Search IDE…. Features Group, interactions, code generations, and acceptance rate per Copilot feature.Use Cases
- See which models and IDEs developers rely on
- Identify languages where Copilot is used most
Clients & Surfaces
What is it?
Requests and tokens per Copilot client and model, and Copilot use outside inline completions: the Copilot CLI, the Copilot app, and the VS Code agent.Metrics Included
Client Requests, Sessions, Total Tokens, Tokens / Request Totals across Copilot clients. Output Share Output tokens as a percentage of all tokens. Requests / Active Day Requests divided by the days with client activity. Token Split Prompt and output tokens per day. Clients and Models Each client, then each model used on it, with the model selection mode, requests, prompt tokens, output tokens, and tokens per request. Surfaces Developers, sessions, requests or messages, and tokens for the Copilot CLI, Copilot app, and VS Code agent.Use Cases
- Understand where Copilot work happens beyond the editor
- Track token consumption per client and model
Top Performers
What is it?
The developers leading on productivity score and on their per-active-day rates.Metrics Included
Top 10 — Productivity Score, Top 10 — Acceptances per Active Day, Top 10 — Lines Added per Active Day Each chart has a Search user… box and View Full List.Use Cases
- Identify power users and Copilot champions
Low Performers
What is it?
The developers with the lowest productivity score and acceptance rate in the period.Metrics Included
Bottom 10 — Productivity Score, Bottom 10 — Acceptance Rate Each chart has a Search user… box and View Full List.Use Cases
- Find developers who may benefit from enablement or training
Pull Requests & Review
What is it?
Pull requests Copilot authored or reviewed, Copilot code review suggestions by comment type, and per-repository activity.Metrics Included
PRs Created, PRs Merged Pull requests created and merged in the period. Reviewed by Copilot Pull requests Copilot reviewed. Copilot-Authored Share Pull requests Copilot created, as a percentage of all pull requests created. Copilot-Reviewed Share Pull requests Copilot reviewed, as a percentage of all pull requests reviewed. Copilot Suggestion Apply Rate Share of Copilot code review suggestions that were applied. Copilot Review Suggestions by Comment Type Suggestions, applied suggestions, and apply rate per comment type. Repositories Created and merged pull requests, and the Copilot-authored and Copilot-reviewed shares, per repository. Search with Search repository….Use Cases
- Measure Copilot’s contribution to code review and delivery
Inactive Users
What is it?
Copilot seats with no recent activity, next to every seat GitHub lists today. This tab follows the Project filter only.Metrics Included
Total Users, Roster Size, Active Users, Inactive Users Seat counts from GitHub’s latest seat list. A seat is inactive when it has no Copilot activity for the inactivity threshold, or was never used. The threshold is 30 days by default and is set per tool in Settings → Inactive User Settings. Inactive Developers Each inactive seat with the user, email, last active date, days inactive, and last editor. Search with Search user…. All Seats Every assigned seat with plan, editor and plugin versions, last activity, last sign-in, and status (Active or Inactive). Search with Search user…. Organization Policies Each organization’s Copilot plan, seat management mode, and IDE Chat, platform Chat, CLI, and public code suggestion policies.Use Cases
- Identify unused licenses to reclaim
- Review Copilot policies per organization
Cost Optimization
What is it?
What GitHub billed for Copilot, including seats and AI credits, and what it bought.Metrics Included
Total Billed Net amount billed, with the gross amount. Acceptances / $1 Accepted suggestions per dollar billed. Cost / Active Developer Net amount billed divided by the developers who used Copilot. Cost / Session Net amount billed divided by Copilot client sessions. AI Credits Used AI credits consumed, with their net amount. Daily Cost ($) Net amount billed per day. Cost per Active Developer Daily cost divided by the developers active that day. AI Credits by Model AI credits used per model. Search with Search model…. Billed Line Items Unit, quantity, gross amount, and net amount per billed SKU.Use Cases
- Track Copilot spend and AI credit consumption
- Compare cost against usage
Cost Savings
What is it?
What the seats cost and what reclaiming inactive seats would save, at the configured seat price. This tab follows the Project filter only.Metrics Included
Total Seats Seats in GitHub’s latest seat list. Cost / Seat / Month The GitHub Copilot seat price from Settings → Tool Cost Settings: the plan you selected there, otherwise the suggested plan. Monthly Cost, Annual Cost Total seats multiplied by the seat price, per month and per year. Inactive Seats Seats with no activity for the inactivity threshold. Monthly / Annual Saving Potential Inactive seats multiplied by the seat price, per month and per year.Use Cases
- Quantify savings from license clean-up
ROI
What is it?
The value Copilot generates, from the hours it saves, against what the seats cost. Calculated with the seat price from Settings → Tool Cost Settings, hours saved per active user per day (0.5 by default), and developer hourly cost ($40 by default). The values in use are shown under ROI & Business Impact Summary. Licences and active users are the average assigned seats and the average seats GitHub counts as active in the period.Metrics Included
Total Hours Saved Active users × hours saved per user per day × 30, per month. Estimated Cost Saved Hours saved multiplied by the developer hourly cost, per month. Monthly Investment Licences multiplied by the seat price. ROI (Cost saved − investment) ÷ investment, as a percentage. ROI Multiplier Cost saved ÷ investment, with the break-even point: the monthly investment divided by the cost saved per day, in days. ROI & Business Impact Summary Investment, hours saved, and cost saved per day, month, and year.Use Cases
- Demonstrate business value
- Justify Copilot investments
Versions & Teams
What is it?
Which IDE, Copilot plugin, and Copilot CLI versions developers use, and Copilot use per GitHub team.Metrics Included
IDE Versions, Copilot Plugin Versions, Copilot CLI Versions Developers per version, from each developer’s latest activity in the period. Teams Members, active developers, adoption, and acceptances for each GitHub team. A developer in several teams counts in each, so team rows are not added together. GitHub reports teams with at least 5 seated members.Use Cases
- Plan IDE and plugin upgrades
- Compare Copilot adoption across teams