Dedicated cores, 2 to 64
x86 runners
Dedicated AMD and Intel runners from 2 to 64 vCPU with local NVMe and unlimited cache.
High clock
Latest-generation AMD EPYC and Intel Xeon, tuned for single-threaded build steps.
Right-sizeable
Six sizes so you can match the machine to the job instead of overpaying by default.
Ubuntu 22.04 and 24.04
Image parity with actions/runner-images.
Most CI jobs are not embarrassingly parallel. A high-clock dedicated core beats a larger count of shared ones for the compile-link-test cycle that dominates real pipelines.
Common questions
- How do I choose a size?
- Start one size above your current runner, check CPU utilisation, and stop when the curve flattens. About 40% of jobs we see average under 30% utilisation, which is money spent on idle cores.
- Are these shared or dedicated cores?
- Dedicated for the duration of the job. Shared vCPU is why the same job takes four minutes on one GitHub-hosted run and nine on the next.
- Is a bigger runner always faster?
- No. A suite that never saturates four cores runs at the same wall-clock time on sixteen and costs four times as much. Check whether you are CPU-bound before scaling up.
Your next build could be twice as fast, at half the price
Start free. Migrating away is the same one line, and we publish that diff too.