GitHub Copilot AI Credits are live: what teams should check now
GitHub Copilot moved to usage-based billing on June 1, 2026, so chat, agents and code review now need budget controls instead of request-count habits.
GitHub Copilot usage-based billing is now active for all Copilot plans. The June 1, 2026 changelog says Copilot plans bill against GitHub AI Credits, while Copilot code review also consumes GitHub Actions minutes. The practical change is simple: inline completions remain bundled, but agent-heavy work now needs spending budgets, usage review and team guidance.
Key takeaways
- GitHub says all Copilot plans now bill based on GitHub AI Credits consumed.
- Copilot code review consumes Actions minutes in addition to AI Credits.
- User-level budgets are generally available for organizations and enterprises.
- Copilot Max is available as an upgrade path for existing Student, Pro and Pro+ subscribers, with new sign-ups still paused at publication time.
- Code completions and Next Edit suggestions remain included in all plans according to GitHub's earlier billing announcement.
Practical LinkLoot angle
Treat this as a FinOps change for coding agents, not just a pricing update. Teams that use Copilot mainly for completions may see little day-to-day difference, while teams running long chat sessions, code review and cloud agent work should measure credit burn by workflow.
| Area | Best use | Cost check | Source |
|---|---|---|---|
| Inline completions | Routine code suggestions and small edits | Confirm they remain outside AI Credit consumption | GitHub announcement |
| Copilot chat and agents | Multi-step repository work, debugging and planning | Track credits by model and session length | GitHub Changelog |
| Copilot code review | Pull request feedback at scale | Count both AI Credits and GitHub Actions minutes | GitHub Changelog |
| Organization budgets | Team-level controls and alerts | Set user or universal budgets before heavy agent adoption | GitHub Changelog |
For LinkLoot readers building agent workflows, the decision point is when to use Copilot's cloud agents versus a cheaper local or API-based coding setup. Keep Copilot for repository-native workflows where GitHub context matters. Use local agents or lower-cost API models for repeatable transforms, bulk refactors and exploratory tasks that do not need GitHub-native review surfaces.
What to verify before you act
Check your plan type first. GitHub's April announcement says annual Pro and Pro+ subscribers keep request-based terms until their annual plan expires, while monthly plans move to the new model on June 1. For organizations, verify whether additional spend is enabled, whether budgets are set at enterprise, cost center or user level, and whether Copilot code review is using the intended Actions runner.
Also verify the model mix. A team that standardizes on high-cost reasoning models for every chat request will spend differently from a team that reserves those models for codebase-wide planning or difficult debugging. The independent 24 AI summary corroborates the central change from premium request units to AI Credits and highlights that chat, autonomous sessions and code review are the usage categories to watch.
Source check
GitHub's June 1 changelog confirms that usage-based billing is live, user-level budgets are generally available and code review consumes both AI Credits and Actions minutes. GitHub's earlier company announcement explains the migration from premium request units to AI Credits, token-based calculation and unchanged base plan prices. 24 AI independently summarizes the same June 1 transition and the plan-level credit framing.
Yes. GitHub says all Copilot plans now bill based on GitHub AI Credits consumed.
For broader tool-stack planning, compare this against the workflows in LinkLoot's AI agent tools guide before standardizing every agent task on one billing model.