$0.005/min
Pay as you go for a 2-vCPU, 8GB runner. Larger hosted shapes scale linearly: $0.010, $0.020, and $0.040 per minute.
Launch pricing
Hosted Linux runners start at $0.005/min with per-second billing after a 10-second minimum. Every job runs in an isolated, ephemeral environment while the execution layer stays portable and provider-neutral.
Pay as you go for a 2-vCPU, 8GB runner. Larger hosted shapes scale linearly: $0.010, $0.020, and $0.040 per minute.
The same 2-vCPU, 8GB runner at the committed team rate. Larger shapes are $0.008, $0.016, and $0.032 per minute.
Hosted control plane plus $0.0015 per 2-vCPU-equivalent minute. Customer-owned compute support remains a roadmap offer, not a private-alpha capability.
Contracted capacity, dedicated cells, regional commitments, and support. Community self-hosting stays free and open source.
Compute is billed per second after a 10-second minimum. There is no recurring hosted free tier: new organizations receive a one-time $10 trial credit, with explicit sponsorships for qualifying open-source projects.
Public network egress is metered separately after the plan allowance. Cache, artifact storage, static egress IPs, and private networking are not bundled into private-alpha compute and are charged only when those capabilities ship.
Admission is manual for trusted GitHub repositories while live reliability and cost evidence accumulates. Published launch rates are the commercial target; no charge begins without explicit enrollment.