Reference
GitHub Actions pricing
By Keith Mazanec, Founder, CostOps ยท Updated February 1, 2026
GitHub bills Actions usage by the minute, with rates that vary by runner type and operating system. This page lists current per-minute rates, OS minute multipliers, and included minutes for each GitHub plan.
Multipliers
Minute multipliers by OS
GitHub applies a multiplier to minutes based on the runner's operating system. These multipliers determine how quickly your included minutes are consumed. For example, 1 minute on a macOS runner uses 10 of your included minutes.
Plans
Included minutes by GitHub plan
Each GitHub plan includes a monthly allowance of Actions minutes (measured in Linux-equivalent minutes). Overages are billed at the per-minute rates below.
| GitHub Plan | Included Minutes |
|---|---|
| Free | 2,000 |
| Pro | 3,000 |
| Team | 3,000 |
| Enterprise | 50,000 |
Rates
Per-minute rates
Rates apply to billable minutes beyond your plan's included allowance. Self-hosted runners are free.
Linux
| Standard (2-core) | $0.0060 |
| 4-core | $0.0120 |
| 8-core | $0.0240 |
| 16-core | $0.0480 |
| 32-core | $0.0960 |
| 64-core | $0.1920 |
| ARM 2-core | $0.0050 |
| ARM 4-core | $0.0100 |
| ARM 8-core | $0.0200 |
| GPU (T4) | $0.0520 |
Windows
| Standard (2-core) | $0.0100 |
| 4-core | $0.0220 |
| 8-core | $0.0440 |
| 16-core | $0.0880 |
| 32-core | $0.1760 |
| 64-core | $0.3520 |
| ARM 2-core | $0.0080 |
| ARM 4-core | $0.0160 |
| ARM 8-core | $0.0320 |
| GPU (T4) | $0.1020 |
macOS
| Standard (3-core) | $0.0620 |
| Large (12-core Intel) | $0.0770 |
| XLarge (M2 Pro + GPU) | $0.1020 |
Rates on this page are sourced from GitHub's billing documentation. CostOps uses these rates to calculate your CI spend from job-level timing data. See the official per-minute rates page for the latest figures.
Related guides
How GitHub Actions Billing Works
Understand the two billing streams, how minutes are metered, and where overages come from.
Overpowered Runners for Lightweight Jobs
Match runner types to job weight and cut spend 50-80%.
Underpowered Runners: Paying More by Going Slow
Right-size runners to cut total cost and build time.
Always-On Self-Hosted Runners
Autoscale and use scale-to-zero to cut self-hosted runner infrastructure cost.