CI Bench

Head to head · CI Bench v1

RunsOn vs Self-hosted runners on Google Cloud

Self-hosted runners on Google Cloud ran jobs 1.2× faster than RunsOn (95% interval 1.2× to 1.3×); RunsOn cost 70% less per job (at 100,000 jobs a month).

Self-hosted options spread fixed monthly costs over the monthly volume, so their cost per job changes with it; see the volume table.

Speed

× faster than GitHub-hosted (ubuntu-latest) · higher is better
RunsOn
1.2×
Self-hosted runners on Google Cloud
1.5×
GitHub-hosted (ubuntu-latest)
1.0×

Cost per job

USD per $1.00 spent on GitHub-hosted (ubuntu-latest), at 100,000 jobs a month · lower is better
RunsOn
$0.12
Self-hosted runners on Google Cloud
$0.41
GitHub-hosted (ubuntu-latest)
$1.00

Workload by workload

RunsOn had the lowest median time of the two on 1 of 12 workloads, Self-hosted runners on Google Cloud on 11.
WorkloadRunsOnSelf-hosted runners on Google CloudLowest median
Burst: 20 jobs at once177 s153 sSelf-hosted runners on Google Cloud
cli327 s310 sSelf-hosted runners on Google Cloud
codex-lint623 s581 sSelf-hosted runners on Google Cloud
container174 s178 sRunsOn
fd89 s64 sSelf-hosted runners on Google Cloud
flask38 s11 sSelf-hosted runners on Google Cloud
guava897 s820 sSelf-hosted runners on Google Cloud
monorepo288 s278 sSelf-hosted runners on Google Cloud
pydantic147 s119 sSelf-hosted runners on Google Cloud
ts570 s533 sSelf-hosted runners on Google Cloud
vite437 s401 sSelf-hosted runners on Google Cloud
vue101 s76 sSelf-hosted runners on Google Cloud

Cost by volume

Mean cost per job in USD, from the benchmark’s sensitivity results
VolumeRunsOnSelf-hosted runners on Google Cloud
100,000 jobs a month$0.121$0.409
10,000 jobs a month$0.250$0.409
1,000 jobs a month$1.431$0.409

Failed runs

Counted, never replaced
RunsOn
1 of 360 measured runs (0.3%)
Self-hosted runners on Google Cloud
1 of 360 measured runs (0.3%)

Each configuration page lists where its failures happened. All losses.

All results for this pair

One line per configuration pairing, generated from the benchmark data