Hosted runners
$0.002per vCPU-minute, Linux x64
- No subscription or platform fee
- Billed per second after a 10-second minimum
- Only customer-code execution is billable
- Seat count never changes the price
Pricing
Pay per second for the code you run.
Planned alpha pricing, subject to change. No subscription or platform fee. Hosted Linux customer-code execution is billed per second after a 10-second minimum.
Self-hosted
$0Apache-2.0 license
- Run the control plane and runners on your own nodes
- Same operation registry, CLI and MCP server
- Source publication follows the release review
- Community support
Customer-owned compute
RoadmapNo published price
- Hosted control plane for your own compute
- Managed dedicated regional cells
- Not part of the alpha
- Tell us your requirements
Runner sizes
One rate per vCPU. Four sizes.
Linux x64, Ubuntu 24.04. Each size includes 4 GB of memory per vCPU and a disposable root disk.
| runs-on label | Size | Rate | Per hour |
|---|---|---|---|
| openrunners-2vcpu-ubuntu-2404 | 2 performance vCPU / 8 GB | $0.004/min | $0.24/hour |
| openrunners-4vcpu-ubuntu-2404 | 4 performance vCPU / 16 GB | $0.008/min | $0.48/hour |
| openrunners-8vcpu-ubuntu-2404 | 8 performance vCPU / 32 GB | $0.016/min | $0.96/hour |
| openrunners-16vcpu-ubuntu-2404 | 16 performance vCPU / 64 GB | $0.032/min | $1.92/hour |
Only customer-code execution is billable. Failed admission and failed provisioning before customer code starts are not billed.
There are no recurring free minutes. OpenRunners may offer discretionary one-time access credits, but they are not an entitlement.
Metered add-ons
Charged only when the meter ships.
Artifact storage, cache, egress and static IPs are not unlimited inside the compute rate. Each becomes billable only after its capability and meter are verified.
- Durable artifact or cold-cache storage
- $0.15/GB-month
- Hot rebuildable cache
- $0.50/GB-month per retained copy
- Public egress above the published allowance
- $0.05/GB in base regions
- Dedicated static egress IP
- $25/month
Cost comparison
How the planned rate compares.
Normalized list prices from public pricing pages. OpenRunners is not the cheapest option; see the provider comparisons for trade-offs.
| Provider | x64 $/vCPU-min | Billing unit | Example month | Source |
|---|---|---|---|---|
| OpenRunners | $0.0020 (planned) | Per second, 10-second minimum (planned) | $400Planned rate; billing is disabled in the alpha | docs/product/pricing.md |
| Blacksmith | $0.0020 | Not stated on the pricing page | $400Before the free monthly allowance | blacksmith.sh/pricing |
| Depot | $0.0030 | Per minute, tracked per second, no minimum | $600Plus a $20 or $200 plan that includes minutes | depot.dev/pricing |
| Namespace | $0.0020 / $0.0030 at 4 GB per vCPU (prepaid / overage) | Per minute, 1-minute minimum | $400–$6004 vCPU / 16 GB shape, prepaid to overage | namespace.so/pricing |
| WarpBuild | $0.0020 | Per minute | $400List rate | warpbuild.com/pricing |
| GitHub-hosted runners | $0.0030 (2–4 cores), falling to $0.00253 at 64 cores | Per job, rounded up to the whole minute | $6004-core larger runner, before per-minute rounding | docs.github.com — Actions runner pricing |
| Ubicloud | $0.000625 standard / $0.0010 premium | Per minute | $125–$200 | ubicloud.com/docs/about/pricing |
List prices from each provider's public pricing page, accessed 2026-10-14. Example month: 50,000 job-minutes on 4 vCPU x64 (200,000 vCPU-minutes), before free tiers, credits, plan fees, and per-minute rounding. Prices change; check each source before deciding.
FAQ
Pricing questions.
No. Billing is disabled. These are planned alpha rates, not active charges, and they may change before billing starts. No charge begins without explicit enrollment.
Only customer-code execution: the GitHub job's start-to-completion time, capped by attested VM elapsed time. Queue time, failed admission and failed provisioning before your code starts are not billed.
There are no recurring free minutes. OpenRunners may offer discretionary one-time access credits during the alpha, but they are not an entitlement.
Some providers list lower rates, as the comparison shows. We will not cut the list rate below what measured node cost and utilization support. We compete on isolation, evidence, auditable billing and an open license.
Yes, before launch. After billing starts, price and regional changes are versioned and apply prospectively.
Questions about a workload? Request access and describe it.
Get on the list for the next admission slot.
Admission is manual for trusted GitHub repositories while the release gates close. Sign in with GitHub to request access and we will review your organization when a slot opens.