Comparison / primary-source research
Choose the operating model. Then compare the bill.
A hosted service and a customer-owned fleet solve different operational problems. The right comparison includes the work your team inherits.
| Approach | Compute ownership | Published offering / our proposal | When to consider it |
|---|---|---|---|
| GitHub-hosted | Provider-managed | Integrated workflow service; standard public-repository compute and plan allowances can be free. | Teams prioritizing simplicity or already covered by allowances. |
| Blacksmith | Blacksmith-hosted | Managed GitHub runners, accelerated caching, persistent disk snapshots, and observability. | Teams wanting faster hosted execution with limited infrastructure work. |
| Actuated | Customer infrastructure | Managed ephemeral microVM runners on customer-owned servers. | Teams with suitable servers that want managed isolation and lifecycle. |
| RunsOn | Customer AWS account | Ephemeral AWS runners; annual software licence plus the customer's cloud bill. | Teams already operating in AWS and evaluating direct infrastructure costs. |
| Plain self-hosted | Customer infrastructure | Official GitHub runner; customer handles isolation, patching, capacity, and cleanup. | Teams with in-house operational expertise and acceptable maintenance cost. |
| Tbench Actions | Proposed hybrid pool | Local-first placement, resource protection, trust-scoped caches, and budgeted overflow are the design target. | Prospective pilots with private Linux CI and spare company compute. Execution not generally available. |
Our proposed advantage
Win on the full cost of a successful pipeline.
The hypothesis is that managed customer-owned capacity can save money without making developers become infrastructure operators. It still needs to be demonstrated.
Schedule near useful state.
A warm local cache can matter more than a low advertised minute rate. Account for transfers and cold starts before moving work.
Respect the working machine.
Resource reservations and opt-in participation are product requirements. A lower infrastructure bill should not hide a productivity bill.
Show the negative cases.
A comparison should include maintenance, depreciation, retries, and fees—even when the answer favors another provider.
Source notes
Published Linux x64 2-core rate: $0.006/minute. GitHub rounds each job up to a whole minute.
Self-hosted runner usage is currently free; infrastructure and associated services remain the customer's responsibility.
Listed Linux x64 2-vCPU rate: $0.004/minute. Its headline savings assume faster execution; we have not independently measured that speedup.
Persistent snapshot-backed build state, with branch-protection controls for cache writes.
Managed CI on customer infrastructure using disposable microVMs. A direct alternative to evaluate.
Customer-owned AWS infrastructure with an annual software licence; cloud infrastructure is billed separately.
Start with your actual workload.
Explore the model, then apply for a scoped pilot. No card or compute commitment.