CI Runnerscirunners.org
Head-to-head

Blacksmith vs Luxor Runners

Answer in brief

Blacksmith suits teams on GitHub Actions that want microVM-isolated runners on bare metal with built-in Docker layer caching and CI observability. Luxor Runners suits teams on GitHub Actions that want per-vCPU-minute Linux x64 and Apple Silicon macOS runners with a fixed egress range. Blacksmith is a managed service hosted by the vendor for GitHub Actions, priced at $0.004/min; Luxor Runners is a managed service hosted by the vendor for GitHub Actions, priced at $0.004/min.

By the CI Runners editors · Updated

Key differences between Blacksmith and Luxor Runners

  • Operating systems: Only Blacksmith documents Windows.
  • Isolation: Blacksmith: Firecracker microVM. Luxor Runners: Single-use VM.
  • Largest runner: 32 vCPU / 128 GB (Blacksmith) versus 16 vCPU / 32 GB (Luxor Runners).

What they have in common

  • Both companies are based in the United States.
  • Both are managed services hosted by the vendor.
  • Both run GitHub Actions jobs.
  • Both document Linux and macOS runners.
  • Both document arm64 runners.
  • Both give every job a fresh environment.
  • Neither is open source.

Blacksmith vs Luxor Runners at a glance

AttributeBlacksmithLuxor Runners
What it isManaged GitHub Actions runners in Firecracker microVMs on the vendor’s own bare-metal fleet, with caches kept next to the runners.GitHub Actions runners on bare-metal Linux x64 and Apple Silicon macOS, one single-use VM per job; called Tenki until November 2026.
TypeManaged runner serviceManaged runner service
CI systemsGitHub ActionsGitHub Actions
Operating systemsLinux, macOS, WindowsLinux, macOS
Architecturesx64, arm64x64, arm64
IsolationEach Linux and Windows job runs in an ephemeral Firecracker microVM on Blacksmith’s bare-metal hosts and is destroyed afterwards; the isolation used for macOS runners is not documentedEach job runs in a single-use VM with its own kernel on bare-metal hosts and is destroyed when the job ends; the hypervisor is not named
Fresh environment per jobYesYes
HardwareBlacksmith describes its hosts as bare-metal machines with gaming CPUs chosen for single-core performance and NVMe drives, in ISO 27001 data centers; CPU models are not named, and macOS runners use Apple M4.Linux x64 runners on bare-metal hosts that Luxor describes as “5th Gen AMD Ryzen” and says it owns and operates, with two vCPUs mapped to one physical core; macOS runners on Apple Silicon M4 Pro.
Performance (vendor claim)Blacksmith says most CI jobs run about 2x faster than on GitHub-hosted runners and cache downloads about 4x faster, and that customers report 2x to 40x faster Docker builds (vendor claims)Luxor says its runners are up to 3x faster than GitHub-hosted runners on its own published benchmarks (vendor claim)
Runner sizes2 to 32 vCPU with 8 to 128 GB RAM on Linux x64; 2 to 32 vCPU with 6 to 96 GB on Linux arm64; 2 to 32 vCPU with 7 to 112 GB on Windows x64 (public beta); 6 or 12 vCPU with 24 or 48 GB on macOSLinux x64: 2 vCPU / 4 GB to 16 vCPU / 32 GB, larger on request; macOS: 2 vCPU / 4 GB to 14 vCPU / 38 GB
CachingStandard cache actions (actions/cache and the setup-* actions) are served from a cache in the same data center as the runners, with 25 GB of free storage per repository per week and eviction of entries unused for 7 days. Docker layers persist on NVMe-backed sticky disks through useblacksmith/setup-docker-builder, billed at $0.50/GB/month; the docs also cover Git checkout, container image and Bazel, Go, Turborepo and Nx build caching.A hosted drop-in backend for the GitHub Actions cache (entries expire after 7 days, up to 10 GiB each), pull-through mirrors for npm, APT, Go modules, Cargo, Maven, CocoaPods, Swift packages and Docker Hub, and sticky disks that keep cache directories between Linux jobs. The docs say the Docker Hub mirror does not cache build layers between jobs.
Docker layer cachingYesNot documented
GPUNot documentedNot documented
ConcurrencyBlacksmith documents no limit on concurrent jobs or vCPUs5 concurrent sessions on Starter and 100 on Team, shared between runner jobs and sandbox sessions; jobs above the limit queue
Static IPsYesYes
SSH into running jobsYesNot documented
PricingThe pricing page lists Ubuntu x64 at $0.004/min, Ubuntu ARM at $0.0025/min, Windows x64 at $0.008/min and macOS M4 at $0.08/min; the docs say runners cost half of GitHub’s per-minute rate. Add-ons: Docker layer caching and sticky disks at $0.50/GB/month and static IPs at $100/IP/month. Flex scheduling costs 40% less for jobs that can wait up to 24 hours to start.Linux x64 is $0.002 per vCPU-minute: $0.004/min (2 vCPU / 4 GB), $0.008 (4 vCPU), $0.016 (8 vCPU), $0.032 (16 vCPU). macOS is $0.020 per vCPU-minute, from $0.040/min (2 vCPU) to $0.280/min (14 vCPU). Cache and sticky-disk storage is $0.20 per GiB-month; the Team plan is $250/month billed monthly or $200/month billed yearly.
Free tier3,000 free minutes per organization per month, counted as x64 2-vCPU minutes; larger, ARM, Windows and macOS runners consume them at weighted rates$100 in free credits once a payment method is added on the $0 Starter plan; Team includes $100 of credits per month
Open sourceNoNo
Self-host / your cloudNo / Not documentedNo / Not documented
ComplianceSOC 2 Type 2; GDPR compliant (as stated by Blacksmith); HIPAA described as in progressLuxor says its parent platform has completed a SOC 2 Type 2 examination and that a SOC 2 report for Luxor Agent Cloud is in progress
Regionsus-east (Ashburn), us-west (Phoenix), eu-west (Amsterdam) and eu-central (Frankfurt); changing region requires contacting supportNot documented
Company headquartersSan Francisco, California, United StatesSeattle, Washington, United States
Good fit forTeams on GitHub Actions that want microVM-isolated runners on bare metal with built-in Docker layer caching and CI observabilityTeams on GitHub Actions that want per-vCPU-minute Linux x64 and Apple Silicon macOS runners with a fixed egress range

Verified November 22, 2026 from each provider’s documentation. Performance figures are vendor claims, not our measurements.

Which should you choose?

Reasons to choose Blacksmith

  • You want each Linux or Windows job in its own Firecracker microVM that is destroyed after the job
  • You want Docker layers kept on persistent NVMe-backed disks between runs
  • You need to run many jobs at once: Blacksmith documents no concurrency limit
  • You want to SSH into running jobs and search CI logs from one console

Reasons to choose Luxor Runners

  • You want Linux x64 runners priced at a flat $0.002 per vCPU-minute, billed by the second with queue time not billed
  • You need Apple Silicon macOS runners (M4 Pro) from the same vendor as your Linux runners
  • You need runner traffic to leave from a fixed CIDR, or a dedicated egress IP at $100 per month
  • You want cache directories kept between Linux jobs on sticky disks instead of archived and restored

Trade-offs to check first

  • Blacksmith: Available to GitHub organizations only, not personal repositories
  • Blacksmith: Windows runners are in public beta and do not support Docker Linux containers
  • Luxor Runners: No Windows runners (stated on luxor.ai); Linux arm64 runners are not documented
  • Luxor Runners: Sticky disks do not support macOS runners or container jobs

Run your slowest workflow on both before deciding. See the full Blacksmith profile and Luxor Runners profile, or the complete comparison.

More comparisons

Frequently asked questions

Is Blacksmith or Luxor Runners better?
Neither is better for every pipeline. Blacksmith suits teams on GitHub Actions that want microVM-isolated runners on bare metal with built-in Docker layer caching and CI observability. Luxor Runners suits teams on GitHub Actions that want per-vCPU-minute Linux x64 and Apple Silicon macOS runners with a fixed egress range.
What is the difference between Blacksmith and Luxor Runners?
Only Blacksmith documents Windows. Isolation is Firecracker microVM on Blacksmith and Single-use VM on Luxor Runners. Largest runner: 32 vCPU / 128 GB (Blacksmith) versus 16 vCPU / 32 GB (Luxor Runners).
Which is open source, Blacksmith or Luxor Runners?
Neither is open source.
How do Blacksmith and Luxor Runners isolate jobs?
Blacksmith: Each Linux and Windows job runs in an ephemeral Firecracker microVM on Blacksmith’s bare-metal hosts and is destroyed afterwards; the isolation used for macOS runners is not documented. Luxor Runners: Each job runs in a single-use VM with its own kernel on bare-metal hosts and is destroyed when the job ends; the hypervisor is not named.
How does Blacksmith pricing compare with Luxor Runners?
Blacksmith: The pricing page lists Ubuntu x64 at $0.004/min, Ubuntu ARM at $0.0025/min, Windows x64 at $0.008/min and macOS M4 at $0.08/min; the docs say runners cost half of GitHub’s per-minute rate. Add-ons: Docker layer caching and sticky disks at $0.50/GB/month and static IPs at $100/IP/month. Flex scheduling costs 40% less for jobs that can wait up to 24 hours to start. Luxor Runners: Linux x64 is $0.002 per vCPU-minute: $0.004/min (2 vCPU / 4 GB), $0.008 (4 vCPU), $0.016 (8 vCPU), $0.032 (16 vCPU). macOS is $0.020 per vCPU-minute, from $0.040/min (2 vCPU) to $0.280/min (14 vCPU). Cache and sticky-disk storage is $0.20 per GiB-month; the Team plan is $250/month billed monthly or $200/month billed yearly. Rates change, so check the Blacksmith and Luxor Runners pricing pages before deciding.
Is Blacksmith or Luxor Runners faster?
We have not benchmarked them, and speed depends on the workload. What each provider says: Blacksmith: Blacksmith says most CI jobs run about 2x faster than on GitHub-hosted runners and cache downloads about 4x faster, and that customers report 2x to 40x faster Docker builds (vendor claims). Luxor Runners: Luxor says its runners are up to 3x faster than GitHub-hosted runners on its own published benchmarks (vendor claim).
Do Blacksmith and Luxor Runners cache Docker layers?
Blacksmith: Standard cache actions (actions/cache and the setup-* actions) are served from a cache in the same data center as the runners, with 25 GB of free storage per repository per week and eviction of entries unused for 7 days. Docker layers persist on NVMe-backed sticky disks through useblacksmith/setup-docker-builder, billed at $0.50/GB/month; the docs also cover Git checkout, container image and Bazel, Go, Turborepo and Nx build caching. Luxor Runners: Docker layer caching is not documented.
Where do Blacksmith and Luxor Runners runners run?
Blacksmith is a managed service hosted by the vendor. Luxor Runners is a managed service hosted by the vendor.

Sources

  1. blacksmith.sh/
  2. blacksmith.sh/pricing
  3. blacksmith.sh/security
  4. docs.blacksmith.sh/blacksmith-runners/overview
  5. luxor.ai/
  6. luxor.ai/docs/migrate-to-luxor.md
  7. luxor.ai/blog/tenki-is-now-luxor-agent-cloud
  8. luxor.ai/products/runners.md

Check before you rely on it. Products change quickly, and this page reflects public documentation as of November 22, 2026. Confirm anything that matters in the provider’s own documentation.

Cite this page

CI Runners. (2026). Blacksmith vs Luxor Runners. cirunners.org. Last reviewed November 22, 2026. https://cirunners.org/compare/blacksmith-vs-luxor-runners/