GitHub Actions runners with the best internet value
Ubicloud Linux x64 leads the measured results at 1.1M per USD. 26 of 26 configurations have a score. All value scores use USD compute rates.
Internet value compares the recorded network score with the runner's quoted compute rate. It divides Mbps by cost per vCPU per minute to highlight transfer performance at a normalized price. The ranking includes every configuration with a usable network result and positive rate. Its leaders, counts, table, and structured data are calculated together during the site build.
By TechCompare · Updated
Current leaders
Results imported 2026-10-10. Pricing checked 2026-10-10. Higher is better. All operating systems and architectures are included. Prices and value use USD. RunJob's EUR rate is converted at €1 = $1.1206 (ECB, 2026-10-09).
Calculator
Ranked runner results
| RunnerRunner | InternetInternet score | Internet valueInternet value | Cost/vCPU/minCost/vCPU/minUSD | Max vCPUMax runner vCPU |
|---|---|---|---|---|
| Ubicloud Linux x64 AMD Ryzen 9 7950X3D 16-Core | 1,104.1 | 1.1M/USD | 0.001 | 30 |
| Ubicloud Linux arm64 Ampere Altra Q80-30 | 670.4 | 1.1M/USD | 0.0006 | 30 |
| Blacksmith Linux arm64 ARM64 (model undisclosed) | 616.7 | 493.4K/USD | 0.0013 | 32 |
| Avrea Linux x64 AMD EPYC 4585PX 16-Core | 921.7 | 460.8K/USD | 0.002 | 32 |
| Blacksmith Linux x64 AMD EPYC (model hidden) | 741.9 | 370.9K/USD | 0.002 | 32 |
| RunJob Linux x64 AMD EPYC 7502 32-Core | 382.6 | 284.5K/USD | 0.0013 | 48 |
| Avrea Windows x64 AMD EPYC 4585PX 16-Core | 823 | 205.8K/USD | 0.004 | 16 |
| Blacksmith Windows x64 AMD EPYC 4565P 16-Core | 731.2 | 182.8K/USD | 0.004 | 32 |
| Latchkey Linux x64 AMD EPYC 7R13 | 214.4 | 171.6K/USD | 0.0013 | 16 |
| Namespace Linux x64 AMD EPYC (model hidden) | 187.3 | 124.9K/USD | 0.0015 | 64 |
| BuildPulse Linux arm64 AWS Graviton (gen undisclosed) | 155.6 | 124.5K/USD | 0.0013 | 64 |
| Namespace Linux arm64 AmpereOne family (SKU undisclosed) | 143.6 | 95.7K/USD | 0.0015 | 32 |
| BuildPulse Linux x64 AMD EPYC 9R45 | 159.1 | 79.5K/USD | 0.002 | 64 |
| GitHub Linux x64 AMD EPYC 9V74 80-Core | 209.9 | 70K/USD | 0.003 | 96 |
| StarSling Linux x64 AMD EPYC (model hidden) | 116.3 | 58.2K/USD | 0.002 | 64 |
| GitHub Windows x86 AMD EPYC 7763 64-Core | 262.4 | 52.5K/USD | 0.005 | 96 |
| GitHub Linux arm64 Microsoft Cobalt 100 | 102.9 | 41.2K/USD | 0.0025 | 64 |
| GitHub Linux x86 AMD EPYC 7763 64-Core | 116.3 | 38.8K/USD | 0.003 | 96 |
| GitHub Windows x64 AMD EPYC 7763 64-Core | 134.2 | 26.8K/USD | 0.005 | 96 |
| GitHub Windows arm64 Cobalt 100 | 120.8 | 24.2K/USD | 0.005 | 64 |
| Blacksmith macOS arm64 Apple M4 Pro | 244.5 | 18.3K/USD | 0.0133 | 12 |
| GitHub macOS arm64 Apple M2 Pro | 273.4 | 13.2K/USD | 0.0207 | 5 |
| Namespace macOS arm64 Apple M4 Pro | 79 | 5.3K/USD | 0.015 | 16 |
| GitHub macOS x64 Intel Core i7-8700B | 76.1 | 4.9K/USD | 0.0155 | 12 |
| Avrea Linux arm64 Apple M5 Max | 18.8 | 3.1K/USD | 0.006 | 16 |
| Avrea macOS arm64 Apple M5 Max | 21.2 | 2.1K/USD | 0.01 | 16 |
Single thread CPU: CoreMark iterations/s. Disk and cache: composite MiB/s. Internet: composite Mbps. Value: score ÷ USD cost/vCPU/min. K means thousand, M means million. Higher is better. Cache uses each runner's best tested cache or artifact backend. Limits: maximum runner vCPU. Prices and value use USD. RunJob's EUR rate is converted at €1 = $1.1206 (ECB, 2026-10-09).
Open the full GitHub Actions runner benchmark toolDivide the network score by compute cost
Internet value equals the recorded composite Mbps score divided by quoted cost per vCPU-minute. A hypothetical score of 200 Mbps at $0.001 per vCPU-minute produces 200,000 per USD. It does not multiply by the runner's core count or convert the score into network capacity per core. The raw Mbps column stays beside value so you can distinguish fast transfer from a high ratio caused by a low compute rate.
Keep the USD conversion explicit
The price assumptions match the main tool: quoted list or overage compute rates, GitHub paid private standard pricing, and consumed CPU pricing for RunJob. EUR rates are converted to USD using the dated ECB reference rate shown on the page. A missing network measurement or an unusable compute rate leaves value unscored. Refreshing the results, researched price, or exchange rate changes the next built ranking without editing a winner name in the article.
Why a transfer ratio is not a savings estimate
A runner may reserve several vCPUs while it waits on one connection, and a consumed CPU service can bill that wait differently from a fixed machine. Provider extras, job rounding, and allowances can further change the total. This test does not include a complete application workflow or a measured invoice. Compare billed durations for representative jobs, including checkout, dependency download, build, tests, and remote uploads, before estimating a migration's savings.
Measurement context
The internet numerator uses the available download and upload rates. A two-direction result uses their geometric mean, while a one-direction result uses that speed alone. Protocol 7 preserves speeds through the time limit using actual bytes and elapsed time. Older timeout records also supply estimates where bytes are available. Endpoints and sample coverage differ, so compare matching endpoints, protocols, and transfer limits for a direct network comparison. These details matter when your job primarily uploads, primarily downloads, or waits on small remote requests.
Verdict
Shortlist runners with useful internet value, then measure traffic to your actual registries and services. This ratio does not include every network, storage, or subscription charge, and it cannot predict the bill for a complete job. Absolute throughput still matters when a transfer is on your critical path.
More runner benchmark scenarios
Frequently asked questions
Which runner currently leads for internet value?
Do these runner rankings update automatically?
Are all operating systems and architectures included?
Does a missing result mean a runner scored zero?
Related tools
Data Transfer Calculator
Estimate transfer times for files over USB, WiFi, Ethernet, and more.
Use tool ➜Memory and Storage Throughput Visualizer
Visualize the massive speed difference between CPU cache, RAM, and storage.
Use tool ➜RAM Latency Calculator
Convert DDR3/DDR4/DDR5 timings (CL, tRCD, tRP, tRAS) into true latency in nanoseconds.
Use tool ➜