GitHub Actions runner benchmarks
11 providers, 102 runners, the same jobs on each: price, real builds, CPU, network, disk and cache. AWS CodeBuild and RunsOn run in your own AWS account; the others run your jobs on theirs. Run by RunsOn from an open harness, last on 22 Nov 2026. Where another runner is faster, it says so.
Scorecard
Scorecardevery provider, every categoryx64 · 2 vCPU Share Save image
RunsOn tmpfs leads on build cost, network and cache; Namespace on CPU and disk; Avrea on build time; boat on price.
| Provider | Runs in | Price per minute ↓ lower | Build cost Rust + TypeScript ↓ lower | Build time Rust + TypeScript ↓ lower | CPU PassMark single-thread ↑ higher | Network download from GitHub ↑ higher | Disk random reads per second ↑ higher | Cache 4 GB restore ↓ lower |
|---|---|---|---|---|---|---|---|---|
| RunsOn | your AWS account | $0.00082 | $0.00462 | 5m 24s3 | 4,2343 | 205 MiB/s2 | 12,5105 | 17s4 |
| RunsOntmpfs | your AWS account | $0.00083 | $0.00421 | 4m 50s2 | 3,7975 | 303 MiB/s1 | workspace in RAM, no disk | 7s1 |
| Avrea | vendor's cloud | $0.00406 | $0.01524 | 3m 49s1 | 4,3582 | 97 MiB/s6 | 53,5983 | 12s3 |
| Blacksmith | vendor's cloud | $0.00407 | $0.02195 | 5m 28s4 | 3,7256 | 95 MiB/s7 | 71,4262 | 9s2 |
| Namespace | vendor's cloud | $0.006011 | $0.0239 | 3m 59s | 4,5431 | 163 MiB/s3 | 105,0541 | 21s5 |
| Warpbuild | vendor's cloud | $0.00408 | $0.02816 | 7m 01s5 | 3,9364 | 111 MiB/s5 | 24,4814 | 26s6 |
| Ubicloud | vendor's cloud | $0.00135 | $0.01043 | 8m 21s6 | 2,8667 | 50 MiB/s10 | 6,8897 | 59s8 |
| boat | vendor's cloud | $0.00031 | $0.0035 | 11m 47s | 2,3029 | 72 MiB/s8 | 10,7836 | 1m 32s10 |
| AWS CodeBuild | your AWS account | $0.00509 | TypeScript: failed · out of memory | TypeScript: failed · out of memory | 2,6028 | 155 MiB/s4 | 1,7639 | 52s7 |
| GitHub | GitHub's cloud | $0.006010 | $0.06297 | 10m 29s7 | 2,24310 | 72 MiB/s9 | 6,8688 | 1m 00s9 |
| Sengi | vendor's cloud | $0.00104 | $0.0173 | 5m 46s | 3,705 | 51 MiB/s | 13,893 | no cache result |
| NamespaceApple silicon | vendor's cloud | arm64 only | Rust: arm64 only | Rust: arm64 only | arm64 only | arm64 only | arm64 only | arm64 only |
| StarSling | vendor's cloud | no 2 vCPU x64 runner | no 2 vCPU x64 runner | no 2 vCPU x64 runner | no 2 vCPU x64 runner | no 2 vCPU x64 runner | no 2 vCPU x64 runner | no 2 vCPU x64 runner |
| WarpbuildApple silicon | vendor's cloud | arm64 only | Rust: arm64 only | Rust: arm64 only | arm64 only | arm64 only | arm64 only | arm64 only |
Each figure is the row's best machine of this size for that column, named in the panel below; the small number is its place. Yellow marks a column's leader, ties within 5% included; rose marks a figure 2 times worse than the column's median or more, RunsOn's included. Build cost and time add up Rust + TypeScript (the builds two providers or more finished at this size); RunsOn is priced at EC2 spot, licenses and plan fees are flat and not in a per-minute figure. Runners of another kind have a row and a place of their own: Apple silicon machines, and RunsOn's large-memory instances with the workspace in RAM (tmpfs). A faded figure takes no place: under 3 runs, or a machine of another size. Runs in: your own AWS account your own AWS account (AWS CodeBuild and RunsOn), so jobs, code and secrets never leave it; the provider's cloud the provider's cloud. Rows are in order of average place across the columns: an order, not a score.
Scorecardevery provider, every categoryx64 · 4 vCPU Share Save image
RunsOn tmpfs leads on build cost and cache; Namespace on build time, CPU and disk; boat on price; StarSling on network.
| Provider | Runs in | Price per minute ↓ lower | Build cost all 3 builds ↓ lower | Build time all 3 builds ↓ lower | CPU PassMark single-thread ↑ higher | Network download from GitHub ↑ higher | Disk random reads per second ↑ higher | Cache 4 GB restore ↓ lower |
|---|---|---|---|---|---|---|---|---|
| Avrea | vendor's cloud | $0.00806 | $0.06304 | 7m 53s2 | 4,4012 | 163 MiB/s4 | 49,5504 | 10s2 |
| RunsOn | your AWS account | $0.00122 | $0.04963 | 10m 55s6 | 4,2553 | 186 MiB/s3 | 11,9506 | 16s4 |
| RunsOntmpfs | your AWS account | $0.00153 | $0.01391 | 8m 19s3 | 3,8025 | 159 MiB/s5 | workspace in RAM, no disk | 5s1 |
| Namespace | vendor's cloud | $0.012012 | $0.08087 | 7m 35s1 | 4,4411 | 124 MiB/s6 | 104,6461 | 18s5 |
| StarSling | vendor's cloud | $0.00808 | $0.06795 | 8m 29s4 | 3,2957 | 217 MiB/s1 | 32,0385 | 23s7 |
| Blacksmith | vendor's cloud | $0.00807 | $0.07156 | 8m 57s5 | 3,6146 | 83 MiB/s9 | 82,0002 | 10s3 |
| Warpbuild | vendor's cloud | $0.00809 | $0.09258 | 11m 34s7 | 4,0814 | 99 MiB/s8 | 63,5583 | 21s6 |
| Ubicloud | vendor's cloud | $0.00255 | $0.04262 | 17m 03s8 | 3,1468 | 54 MiB/s11 | 7,8018 | 56s10 |
| boat | vendor's cloud | $0.00061 | Docker: not run: needs more than 8 GB of memory | Docker: not run: needs more than 8 GB of memory | 3,0159 | 61 MiB/s10 | 10,3387 | 1m 36s11 |
| AWS CodeBuild | your AWS account | $0.010010 | TypeScript: failed · out of memory | TypeScript: failed · out of memory | 2,62110 | 213 MiB/s2 | 1,79410 | 35s8 |
| GitHub | GitHub's cloud | $0.012011 | $0.24329 | 20m 16s9 | 2,36511 | 100 MiB/s7 | 7,3329 | 38s9 |
| Sengi | vendor's cloud | $0.00204 | $0.0614 | 10m 14s | 3,724 | 82 MiB/s | 13,632 | 7s |
| NamespaceApple silicon | vendor's cloud | arm64 only | Rust: arm64 only | Rust: arm64 only | arm64 only | arm64 only | arm64 only | arm64 only |
| WarpbuildApple silicon | vendor's cloud | arm64 only | Rust: arm64 only | Rust: arm64 only | arm64 only | arm64 only | arm64 only | arm64 only |
Each figure is the row's best machine of this size for that column, named in the panel below; the small number is its place. Yellow marks a column's leader, ties within 5% included; rose marks a figure 2 times worse than the column's median or more, RunsOn's included. Build cost and time add up all 3 builds; RunsOn is priced at EC2 spot, licenses and plan fees are flat and not in a per-minute figure. Runners of another kind have a row and a place of their own: Apple silicon machines, and RunsOn's large-memory instances with the workspace in RAM (tmpfs). A faded figure takes no place: under 3 runs, or a machine of another size. Runs in: your own AWS account your own AWS account (AWS CodeBuild and RunsOn), so jobs, code and secrets never leave it; the provider's cloud the provider's cloud. Rows are in order of average place across the columns: an order, not a score.
Scorecardevery provider, every categoryx64 · 8 vCPU Share Save image
RunsOn leads on network; RunsOn tmpfs on cache; Namespace on build time, CPU and disk; boat on price and build cost.
| Provider | Runs in | Price per minute ↓ lower | Build cost all 3 builds ↓ lower | Build time all 3 builds ↓ lower | CPU PassMark single-thread ↑ higher | Network download from GitHub ↑ higher | Disk random reads per second ↑ higher | Cache 4 GB restore ↓ lower |
|---|---|---|---|---|---|---|---|---|
| Avrea | vendor's cloud | $0.01606 | $0.09476 | 5m 55s2 | 4,4342 | 150 MiB/s4 | 52,1213 | 10s3 |
| RunsOntmpfs | your AWS account | $0.00243 | $0.01712 | 6m 16s4 | 3,7984 | 172 MiB/s2 | workspace in RAM, no disk | 5s1 |
| RunsOn | your AWS account | $0.00242 | $0.03353 | 7m 15s6 | 3,7405 | 178 MiB/s1 | 11,0837 | 11s4 |
| Namespace | vendor's cloud | $0.024012 | $0.07584 | 5m 43s1 | 4,5621 | 137 MiB/s5 | 106,2281 | 18s5 |
| Blacksmith | vendor's cloud | $0.01607 | $0.09927 | 6m 12s3 | 3,6656 | 105 MiB/s8 | 74,0952 | 8s2 |
| StarSling | vendor's cloud | $0.01608 | $0.10328 | 6m 27s5 | 3,4128 | 130 MiB/s6 | 52,0584 | 24s7 |
| boat | vendor's cloud | $0.00121 | $0.01321 | 11m 00s9 | 2,8809 | 79 MiB/s10 | 11,3836 | 1m 11s11 |
| Warpbuild | vendor's cloud | $0.01609 | $0.12609 | 7m 53s7 | 4,0403 | 101 MiB/s9 | 37,1535 | 23s6 |
| Ubicloud | vendor's cloud | $0.00505 | $0.07615 | 9m 31s8 | 3,6167 | 50 MiB/s11 | 9,5438 | 55s10 |
| AWS CodeBuild | your AWS account | $0.020010 | $0.342711 | 17m 08s11 | 2,62610 | 161 MiB/s3 | 1,75810 | 35s9 |
| GitHub | GitHub's cloud | $0.022011 | $0.298510 | 13m 34s10 | 2,35011 | 107 MiB/s7 | 7,5599 | 33s8 |
| Sengi | vendor's cloud | $0.00404 | $0.0892 | 7m 26s | 3,745 | 56 MiB/s | 13,912 | 6s |
| NamespaceApple silicon | vendor's cloud | arm64 only | Rust: arm64 only | Rust: arm64 only | arm64 only | arm64 only | arm64 only | arm64 only |
| WarpbuildApple silicon | vendor's cloud | arm64 only | Rust: arm64 only | Rust: arm64 only | arm64 only | arm64 only | arm64 only | arm64 only |
Each figure is the row's best machine of this size for that column, named in the panel below; the small number is its place. Yellow marks a column's leader, ties within 5% included; rose marks a figure 2 times worse than the column's median or more, RunsOn's included. Build cost and time add up all 3 builds; RunsOn is priced at EC2 spot, licenses and plan fees are flat and not in a per-minute figure. Runners of another kind have a row and a place of their own: Apple silicon machines, and RunsOn's large-memory instances with the workspace in RAM (tmpfs). A faded figure takes no place: under 3 runs, or a machine of another size. Runs in: your own AWS account your own AWS account (AWS CodeBuild and RunsOn), so jobs, code and secrets never leave it; the provider's cloud the provider's cloud. Rows are in order of average place across the columns: an order, not a score.
Scorecardevery provider, every categoryarm64 · 2 vCPU Share Save image
RunsOn leads on price and build cost; RunsOn tmpfs on network and cache; Avrea on build time and CPU; Blacksmith on disk.
| Provider | Runs in | Price per minute ↓ lower | Build cost Rust + TypeScript ↓ lower | Build time Rust + TypeScript ↓ lower | CPU PassMark single-thread ↑ higher | Network download from GitHub ↑ higher | Disk random reads per second ↑ higher | Cache 4 GB restore ↓ lower |
|---|---|---|---|---|---|---|---|---|
| RunsOn | your AWS account | $0.00081 | $0.00451 | 5m 06s3 | 2,4753 | 177 MiB/s2 | 13,5335 | 19s3 |
| RunsOntmpfs | your AWS account | $0.00102 | $0.00552 | 5m 03s2 | 2,4664 | 191 MiB/s1 | workspace in RAM, no disk | 8s1 |
| Avrea | vendor's cloud | $0.01208 | $0.04076 | 3m 24s1 | 4,7441 | 141 MiB/s5 | 20,9092 | 12s2 |
| Blacksmith | vendor's cloud | $0.00254 | $0.03045 | 12m 09s6 | 1,3307 | 92 MiB/s8 | 31,0281 | 22s4 |
| Warpbuild | vendor's cloud | $0.00305 | $0.02604 | 8m 39s4 | 1,9395 | 167 MiB/s3 | 1,7798 | 28s6 |
| NamespaceApple silicon | vendor's cloud | $0.04209 | $0.1333 | 3m 11s | 4,5942 | 143 MiB/s4 | 19,8293 | 24s5 |
| Ubicloud | vendor's cloud | $0.00133 | $0.01763 | 14m 07s7 | 1,2978 | 58 MiB/s9 | 8,4526 | 1m 04s9 |
| GitHub | GitHub's cloud | $0.00506 | $0.04577 | 9m 08s5 | 1,8696 | 120 MiB/s7 | 8,1197 | 53s8 |
| Namespace | vendor's cloud | $0.00607 | TypeScript: failed · out of memory | TypeScript: failed · out of memory | 1,1469 | 122 MiB/s6 | 15,9414 | 43s7 |
| AWS CodeBuild | your AWS account | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner |
| boat | vendor's cloud | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner |
| Sengi | vendor's cloud | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner |
| StarSling | vendor's cloud | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner |
| WarpbuildApple silicon | vendor's cloud | no 2 vCPU arm64 runner | no 2 vCPU arm64 runner | no 2 vCPU arm64 runner | no 2 vCPU arm64 runner | no 2 vCPU arm64 runner | no 2 vCPU arm64 runner | no 2 vCPU arm64 runner |
Each figure is the row's best machine of this size for that column, named in the panel below; the small number is its place. Yellow marks a column's leader, ties within 5% included; rose marks a figure 2 times worse than the column's median or more, RunsOn's included. Build cost and time add up Rust + TypeScript (the builds two providers or more finished at this size); RunsOn is priced at EC2 spot, licenses and plan fees are flat and not in a per-minute figure. Runners of another kind have a row and a place of their own: Apple silicon machines, and RunsOn's large-memory instances with the workspace in RAM (tmpfs). A faded figure takes no place: under 3 runs, or a machine of another size. Runs in: your own AWS account your own AWS account (AWS CodeBuild and RunsOn), so jobs, code and secrets never leave it; the provider's cloud the provider's cloud. Rows are in order of average place across the columns: an order, not a score.
Scorecardevery provider, every categoryarm64 · 4 vCPU Share Save image
RunsOn leads on price, build cost and network; RunsOn tmpfs on cache; Avrea on CPU; Blacksmith on disk; Namespace Apple silicon on build time.
| Provider | Runs in | Price per minute ↓ lower | Build cost all 3 builds ↓ lower | Build time all 3 builds ↓ lower | CPU PassMark single-thread ↑ higher | Network download from GitHub ↑ higher | Disk random reads per second ↑ higher | Cache 4 GB restore ↓ lower |
|---|---|---|---|---|---|---|---|---|
| RunsOn | your AWS account | $0.00131 | $0.01461 | 9m 45s4 | 2,4723 | 172 MiB/s1 | 13,3685 | 16s3 |
| RunsOntmpfs | your AWS account | $0.00172 | $0.01932 | 8m 47s3 | 2,4674 | 167 MiB/s2 | workspace in RAM, no disk | 5s1 |
| Avrea | vendor's cloud | $0.02408 | $0.18047 | 7m 31s2 | 4,8161 | 86 MiB/s8 | 20,3012 | 10s2 |
| Blacksmith | vendor's cloud | $0.00504 | $0.08604 | 17m 12s6 | 1,3407 | 97 MiB/s7 | 32,6911 | 20s4 |
| NamespaceApple silicon | vendor's cloud | $0.08409 | $0.60979 | 7m 16s1 | 4,5842 | 127 MiB/s4 | 18,5703 | 21s5 |
| Warpbuild | vendor's cloud | $0.00605 | $0.09635 | 16m 03s5 | 1,9405 | 158 MiB/s3 | 1,8128 | 27s6 |
| GitHub | GitHub's cloud | $0.00806 | $0.13876 | 17m 21s7 | 1,8706 | 118 MiB/s6 | 7,6947 | 37s7 |
| Ubicloud | vendor's cloud | $0.00253 | $0.05803 | 23m 13s8 | 1,2868 | 74 MiB/s9 | 8,3346 | 1m 01s9 |
| Namespace | vendor's cloud | $0.01207 | $0.24418 | 23m 26s9 | 1,1629 | 125 MiB/s5 | 16,7594 | 39s8 |
| AWS CodeBuild | your AWS account | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner |
| boat | vendor's cloud | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner |
| Sengi | vendor's cloud | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner |
| StarSling | vendor's cloud | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner |
| WarpbuildApple · 6 vCPU | vendor's cloud | $0.0800 | $0.7293 | 9m 07s | 3,006 | 68 MiB/s | 29,080 | 37s |
Each figure is the row's best machine of this size for that column, named in the panel below; the small number is its place. Yellow marks a column's leader, ties within 5% included; rose marks a figure 2 times worse than the column's median or more, RunsOn's included. Build cost and time add up all 3 builds; RunsOn is priced at EC2 spot, licenses and plan fees are flat and not in a per-minute figure. Runners of another kind have a row and a place of their own: Apple silicon machines, and RunsOn's large-memory instances with the workspace in RAM (tmpfs). A faded figure takes no place: under 3 runs, or a machine of another size. Runs in: your own AWS account your own AWS account (AWS CodeBuild and RunsOn), so jobs, code and secrets never leave it; the provider's cloud the provider's cloud. Rows are in order of average place across the columns: an order, not a score.
Scorecardevery provider, every categoryarm64 · 8 vCPU Share Save image
RunsOn leads on price; RunsOn tmpfs on build cost, network and cache; Avrea on build time and CPU; Blacksmith on disk.
| Provider | Runs in | Price per minute ↓ lower | Build cost all 3 builds ↓ lower | Build time all 3 builds ↓ lower | CPU PassMark single-thread ↑ higher | Network download from GitHub ↑ higher | Disk random reads per second ↑ higher | Cache 4 GB restore ↓ lower |
|---|---|---|---|---|---|---|---|---|
| RunsOntmpfs | your AWS account | $0.00222 | $0.01611 | 6m 31s2 | 2,4713 | 253 MiB/s1 | workspace in RAM, no disk | 4s1 |
| RunsOn | your AWS account | $0.00221 | $0.02332 | 6m 53s3 | 2,4722 | 161 MiB/s3 | 13,4624 | 10s3 |
| Avrea | vendor's cloud | $0.04807 | $0.27567 | 5m 45s1 | 4,8151 | 84 MiB/s6 | 20,3972 | 10s2 |
| Blacksmith | vendor's cloud | $0.01004 | $0.11964 | 11m 58s5 | 1,3445 | 101 MiB/s5 | 31,9051 | 17s4 |
| Warpbuild | vendor's cloud | $0.01205 | $0.12245 | 10m 12s4 | 1,9404 | 198 MiB/s2 | 1,7816 | 21s5 |
| Namespace | vendor's cloud | $0.02406 | $0.25496 | 17m 10s6 | 1,1667 | 134 MiB/s4 | 17,4523 | 34s6 |
| Ubicloud | vendor's cloud | $0.00503 | $0.08603 | 17m 12s7 | 1,2986 | 57 MiB/s7 | 8,4535 | 50s7 |
| AWS CodeBuild | your AWS account | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner |
| boat | vendor's cloud | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner |
| GitHub | GitHub's cloud | no 8 vCPU arm64 runner | no 8 vCPU arm64 runner | no 8 vCPU arm64 runner | no 8 vCPU arm64 runner | no 8 vCPU arm64 runner | no 8 vCPU arm64 runner | no 8 vCPU arm64 runner |
| NamespaceApple silicon | vendor's cloud | no 8 vCPU arm64 runner | no 8 vCPU arm64 runner | no 8 vCPU arm64 runner | no 8 vCPU arm64 runner | no 8 vCPU arm64 runner | no 8 vCPU arm64 runner | no 8 vCPU arm64 runner |
| Sengi | vendor's cloud | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner |
| StarSling | vendor's cloud | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner | no arm64 runner |
| WarpbuildApple · 12 vCPU | vendor's cloud | $0.1600 | $1.15 | 7m 13s | 3,145 | 70 MiB/s | 30,102 | 1m 06s |
Each figure is the row's best machine of this size for that column, named in the panel below; the small number is its place. Yellow marks a column's leader, ties within 5% included; rose marks a figure 2 times worse than the column's median or more, RunsOn's included. Build cost and time add up all 3 builds; RunsOn is priced at EC2 spot, licenses and plan fees are flat and not in a per-minute figure. Runners of another kind have a row and a place of their own: Apple silicon machines, and RunsOn's large-memory instances with the workspace in RAM (tmpfs). A faded figure takes no place: under 3 runs, or a machine of another size. Runs in: your own AWS account your own AWS account (AWS CodeBuild and RunsOn), so jobs, code and secrets never leave it; the provider's cloud the provider's cloud. Rows are in order of average place across the columns: an order, not a score.
Real builds
Real buildstime and cost per buildlower is betterx64 · 2 vCPU Share Save image
Rust cold build of spotify-player: cargo test and clippy
boat is the best value and the cheapest; Avrea is the fastest.
- boat 4 GB 4m 35s $0.0014
- RunsOn m8a.large · EBS gp3 · 8 GB 3m 11s $0.0026
- RunsOntmpfs r8a.large · 16 GB 3m 07s $0.0027
- Avrea 8 GB 2m 25s $0.0096
- Ubicloud standard · 8 GB 4m 49s $0.0060
- Namespace 2x8 · 8 GB 2m 36s $0.0156
- Blacksmith 8 GB 3m 33s $0.0142
- Warpbuild 8 GB 4m 10s $0.0167
- GitHub 8 GB 5m 53s $0.0353
- AWS CodeBuild small · 3 GB 6m 52s $0.0343
- Sengi premium · 8 GB 3m 15s $0.0097
- NamespaceApple silicon arm64 only
- StarSling no 2 vCPU x64 runner
- WarpbuildApple silicon arm64 only
TypeScript cold bun install and typecheck of opencode
RunsOn is the best value; RunsOn tmpfs is the cheapest; Avrea is the fastest.
- RunsOn m8a.large · EBS gp3 · 8 GB 2m 13s $0.0020
- RunsOntmpfs r8a.large · 16 GB 1m 44s $0.0015
- Avrea 8 GB 1m 24s $0.0056
- Ubicloud standard · 8 GB 3m 33s $0.0044
- Blacksmith 8 GB 1m 56s $0.0077
- Warpbuild 8 GB 2m 51s $0.0114
- GitHub 8 GB 4m 36s $0.0276
- Namespace 2x8 · 8 GB 1m 23s $0.0083
- boat 4 GB 7m 12s $0.0022
- Sengi premium · 8 GB 2m 31s $0.0076
- AWS CodeBuild failed · out of memory
- NamespaceApple silicon arm64 only
- StarSling no 2 vCPU x64 runner
- WarpbuildApple silicon arm64 only
Docker PostHog's production image, built cold with buildx
Only RunsOn finished this build.
- RunsOn r8a.large · EBS gp3 · 16 GB 12m 30s $0.0111
- Avrea not run: needs more than 8 GB of memory
- AWS CodeBuild not run: needs more than 8 GB of memory
- Blacksmith not run: needs more than 8 GB of memory
- boat not run: needs more than 8 GB of memory
- GitHub not run: needs more than 8 GB of memory
- Namespace not run: needs more than 8 GB of memory
- NamespaceApple silicon arm64 only
- RunsOntmpfs not run: needs more than 16 GB of RAM with tmpfs, where Docker's data root lives in RAM
- Sengi not run: needs more than 8 GB of memory
- StarSling no 2 vCPU x64 runner
- Ubicloud not run: needs more than 8 GB of memory
- Warpbuild not run: needs more than 8 GB of memory
- WarpbuildApple silicon arm64 only
Each row's best-value machine of this size for that build (lowest time × cost, ranked on p90), in that order, named under the provider. Time is the median whole job, cost is that job at the provider's per-minute price (RunsOn at EC2 spot). Yellow marks the fastest and the cheapest. Apple silicon and tmpfs (workspace in RAM) machines are rows of their own; a faded row is provisional or a machine of another size, and takes no place. A row that did not finish a build stays listed with the reason.
Full build benchmark every phase, p90 and worst runs, queue under a burst, run-to-run spread, bigger runners
Real buildstime and cost per buildlower is betterx64 · 4 vCPU Share Save image
Rust cold build of spotify-player: cargo test and clippy
RunsOn is the best value; boat is the cheapest; Avrea is the fastest.
- RunsOn c8a.xlarge · EBS gp3 · 8 GB 2m 03s $0.0035
- RunsOntmpfs r8a.xlarge · 32 GB 2m 03s $0.0035
- Namespace 4x8 · 8 GB 1m 42s $0.0102
- Avrea 16 GB 1m 37s $0.0129
- StarSling 16 GB 1m 41s $0.0135
- Ubicloud standard · 16 GB 3m 16s $0.0082
- boat 8 GB 2m 42s $0.0016
- Warpbuild 16 GB 2m 36s $0.0208
- Blacksmith 16 GB 2m 00s $0.0159
- AWS CodeBuild medium · 8 GB 3m 46s $0.0377
- GitHub 16 GB 3m 48s $0.0456
- Sengi premium · 16 GB 1m 58s $0.0118
- NamespaceApple silicon arm64 only
- WarpbuildApple silicon arm64 only
TypeScript cold bun install and typecheck of opencode
RunsOn tmpfs is the best value; boat is the cheapest; Avrea is the fastest.
- RunsOntmpfs r8a.xlarge · 32 GB 1m 01s $0.0020
- boat 8 GB 1m 44s $0.0010
- RunsOn m8azn.xlarge · gp3 1000 MiB/s · 16 GB 59s $0.0060
- Avrea 16 GB 50s $0.0066
- StarSling 16 GB 55s $0.0073
- Ubicloud standard · 16 GB 1m 52s $0.0046
- Namespace 4x16 · 16 GB 56s $0.0112
- Warpbuild 16 GB 1m 20s $0.0107
- Blacksmith 16 GB 1m 09s $0.0091
- GitHub 16 GB 2m 00s $0.0240
- Sengi premium · 16 GB 1m 18s $0.0078
- AWS CodeBuild failed · out of memory
- NamespaceApple silicon arm64 only
- WarpbuildApple silicon arm64 only
Docker PostHog's production image, built cold with buildx
RunsOn tmpfs is the best value and the cheapest; Namespace is the fastest.
- RunsOntmpfs r8a.xlarge · 32 GB 5m 15s $0.0084
- Avrea 16 GB 5m 26s $0.0435
- StarSling 16 GB 5m 53s $0.0471
- Namespace 4x16 · 16 GB 4m 57s $0.0594
- RunsOn m8azn.xlarge · gp3 1000 MiB/s · 16 GB 7m 54s $0.0401
- Blacksmith 16 GB 5m 49s $0.0465
- Ubicloud standard · 16 GB 11m 56s $0.0298
- Warpbuild 16 GB 7m 38s $0.0611
- GitHub 16 GB 14m 28s $0.1736
- Sengi premium · 16 GB 6m 58s $0.0418
- AWS CodeBuild not run: needs more than 8 GB of memory
- boat not run: needs more than 8 GB of memory
- NamespaceApple silicon arm64 only
- WarpbuildApple silicon arm64 only
Each row's best-value machine of this size for that build (lowest time × cost, ranked on p90), in that order, named under the provider. Time is the median whole job, cost is that job at the provider's per-minute price (RunsOn at EC2 spot). Yellow marks the fastest and the cheapest. Apple silicon and tmpfs (workspace in RAM) machines are rows of their own; a faded row is provisional or a machine of another size, and takes no place. A row that did not finish a build stays listed with the reason.
Full build benchmark every phase, p90 and worst runs, queue under a burst, run-to-run spread, bigger runners
Real buildstime and cost per buildlower is betterx64 · 8 vCPU Share Save image
Rust cold build of spotify-player: cargo test and clippy
RunsOn tmpfs is the best value; boat is the cheapest; StarSling is the fastest.
- RunsOntmpfs m8a.2xlarge · 32 GB 1m 28s $0.0041
- RunsOn m8a.2xlarge · EBS gp3 · 32 GB 1m 28s $0.0044
- Namespace 8x16 · 16 GB 1m 14s $0.0147
- Avrea 32 GB 1m 13s $0.0195
- StarSling 32 GB 1m 12s $0.0192
- boat 16 GB 1m 39s $0.0020
- Ubicloud premium · 32 GB 1m 56s $0.0154
- Warpbuild 32 GB 1m 36s $0.0255
- Blacksmith 32 GB 1m 22s $0.0217
- AWS CodeBuild large · 16 GB 2m 34s $0.0512
- GitHub 32 GB 2m 32s $0.0555
- Sengi premium · 32 GB 1m 20s $0.0160
- NamespaceApple silicon arm64 only
- WarpbuildApple silicon arm64 only
TypeScript cold bun install and typecheck of opencode
RunsOn tmpfs is the best value; boat is the cheapest; Namespace is the fastest.
- RunsOntmpfs m8a.2xlarge · 32 GB 48s $0.0025
- RunsOn m8a.2xlarge · gp3 1000 MiB/s · 32 GB 47s $0.0048
- boat 16 GB 51s $0.0010
- Avrea 32 GB 37s $0.0097
- StarSling 32 GB 44s $0.0117
- Ubicloud premium · 32 GB 1m 02s $0.0082
- Blacksmith 32 GB 44s $0.0117
- Warpbuild 32 GB 55s $0.0147
- Namespace 8x32 · 32 GB 36s $0.0144
- AWS CodeBuild large · 16 GB 1m 42s $0.0338
- GitHub 32 GB 1m 20s $0.0291
- Sengi premium · 32 GB 42s $0.0084
- NamespaceApple silicon arm64 only
- WarpbuildApple silicon arm64 only
Docker PostHog's production image, built cold with buildx
RunsOn tmpfs is the best value; boat is the cheapest; Namespace is the fastest.
- RunsOntmpfs m8a.2xlarge · 32 GB 4m 00s $0.0105
- RunsOn m8a.2xlarge · gp3 1000 MiB/s · 32 GB 5m 00s $0.0243
- Namespace 8x16 · 16 GB 3m 54s $0.0467
- boat 16 GB 8m 31s $0.0102
- Avrea 32 GB 4m 06s $0.0655
- Blacksmith 32 GB 4m 07s $0.0657
- StarSling 32 GB 4m 31s $0.0723
- Warpbuild 32 GB 5m 22s $0.0859
- Ubicloud premium · 32 GB 6m 34s $0.0525
- GitHub 32 GB 9m 43s $0.2138
- AWS CodeBuild large · 16 GB 12m 53s $0.2577
- Sengi premium · 32 GB 5m 24s $0.0648
- NamespaceApple silicon arm64 only
- WarpbuildApple silicon arm64 only
Each row's best-value machine of this size for that build (lowest time × cost, ranked on p90), in that order, named under the provider. Time is the median whole job, cost is that job at the provider's per-minute price (RunsOn at EC2 spot). Yellow marks the fastest and the cheapest. Apple silicon and tmpfs (workspace in RAM) machines are rows of their own; a faded row is provisional or a machine of another size, and takes no place. A row that did not finish a build stays listed with the reason.
Full build benchmark every phase, p90 and worst runs, queue under a burst, run-to-run spread, bigger runners
Real buildstime and cost per buildlower is betterarm64 · 2 vCPU Share Save image
Rust cold build of spotify-player: cargo test and clippy
RunsOn is the best value and the cheapest; Avrea is the fastest.
- RunsOn m9gd.large · local NVMe · 8 GB 3m 03s $0.0026
- RunsOntmpfs r9g.large · 16 GB 3m 24s $0.0037
- Avrea 8 GB 2m 03s $0.0245
- Warpbuild 8 GB 4m 59s $0.0149
- Blacksmith 6 GB 6m 41s $0.0167
- Ubicloud standard · 8 GB 9m 06s $0.0114
- GitHub 8 GB 5m 42s $0.0285
- NamespaceApple silicon 2x8 · 8 GB 2m 03s $0.0857
- Namespace 2x8 · 8 GB 10m 08s $0.0608
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
- WarpbuildApple silicon no 2 vCPU arm64 runner
TypeScript cold bun install and typecheck of opencode
RunsOn is the cheapest; RunsOn tmpfs is the best value; Avrea is the fastest.
- RunsOntmpfs r9g.large · 16 GB 1m 39s $0.0019
- RunsOn m9gd.large · local NVMe · 8 GB 2m 04s $0.0018
- Avrea 8 GB 1m 21s $0.0162
- Ubicloud standard · 8 GB 5m 01s $0.0063
- Blacksmith 6 GB 5m 28s $0.0136
- GitHub 8 GB 3m 27s $0.0172
- Warpbuild 8 GB 3m 40s $0.0110
- NamespaceApple silicon 2x8 · 8 GB 1m 08s $0.0476
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- Namespace failed · out of memory
- Sengi no arm64 runner
- StarSling no arm64 runner
- WarpbuildApple silicon no 2 vCPU arm64 runner
Docker PostHog's production image, built cold with buildx
No 2 vCPU arm64 runner finished this build.
- Avrea not run: needs more than 8 GB of memory
- AWS CodeBuild no arm64 runner
- Blacksmith not run: needs more than 8 GB of memory
- boat no arm64 runner
- GitHub not run: needs more than 8 GB of memory
- Namespace not run: needs more than 8 GB of memory
- NamespaceApple silicon not run: needs more than 8 GB of memory
- RunsOn not run: needs more than 8 GB of memory
- RunsOntmpfs not run: needs more than 16 GB of RAM with tmpfs, where Docker's data root lives in RAM
- Sengi no arm64 runner
- StarSling no arm64 runner
- Ubicloud not run: needs more than 8 GB of memory
- Warpbuild not run: needs more than 8 GB of memory
- WarpbuildApple silicon no 2 vCPU arm64 runner
Each row's best-value machine of this size for that build (lowest time × cost, ranked on p90), in that order, named under the provider. Time is the median whole job, cost is that job at the provider's per-minute price (RunsOn at EC2 spot). Yellow marks the fastest and the cheapest. Apple silicon and tmpfs (workspace in RAM) machines are rows of their own; a faded row is provisional or a machine of another size, and takes no place. A row that did not finish a build stays listed with the reason.
Full build benchmark every phase, p90 and worst runs, queue under a burst, run-to-run spread, bigger runners
Real buildstime and cost per buildlower is betterarm64 · 4 vCPU Share Save image
Rust cold build of spotify-player: cargo test and clippy
RunsOn is the best value and the cheapest; Namespace Apple silicon is the fastest.
- RunsOn m9g.xlarge · EBS gp3 · 16 GB 2m 01s $0.0039
- RunsOntmpfs m9g.xlarge · 16 GB 2m 05s $0.0039
- Warpbuild 16 GB 3m 09s $0.0189
- Avrea 16 GB 1m 29s $0.0356
- Blacksmith 16 GB 4m 07s $0.0205
- GitHub 16 GB 3m 27s $0.0276
- Ubicloud standard · 16 GB 5m 47s $0.0145
- NamespaceApple silicon 4x16 · 16 GB 1m 20s $0.1120
- Namespace 4x8 · 8 GB 6m 12s $0.0372
- WarpbuildApple silicon 6 vCPU · xfast (Apple M4 Pro) · 14 GB 1m 21s $0.1080
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
TypeScript cold bun install and typecheck of opencode
RunsOn is the best value and the cheapest; Namespace Apple silicon is the fastest.
- RunsOn m9gd.xlarge · local NVMe · 16 GB 1m 03s $0.0017
- RunsOntmpfs m9g.xlarge · 16 GB 1m 02s $0.0022
- Warpbuild 16 GB 1m 36s $0.0095
- Ubicloud standard · 16 GB 2m 34s $0.0064
- GitHub 16 GB 1m 42s $0.0136
- Blacksmith 16 GB 2m 12s $0.0110
- Avrea 16 GB 48s $0.0190
- NamespaceApple silicon 4x16 · 16 GB 47s $0.0658
- Namespace 4x16 · 16 GB 3m 08s $0.0376
- WarpbuildApple silicon 6 vCPU · xfast (Apple M4 Pro) · 14 GB 53s $0.0707
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
Docker PostHog's production image, built cold with buildx
RunsOn is the cheapest; RunsOn tmpfs is the best value; Namespace Apple silicon is the fastest.
- RunsOntmpfs r9g.xlarge · 32 GB 5m 40s $0.0132
- RunsOn m9gd.xlarge · local NVMe · 16 GB 6m 41s $0.0090
- Blacksmith 16 GB 10m 54s $0.0545
- Ubicloud standard · 16 GB 14m 52s $0.0372
- Avrea 16 GB 5m 15s $0.1258
- Warpbuild 16 GB 11m 18s $0.0678
- GitHub 16 GB 12m 12s $0.0975
- Namespace 4x16 · 16 GB 14m 07s $0.1693
- NamespaceApple silicon 4x16 · 16 GB 5m 09s $0.4319
- WarpbuildApple silicon 6 vCPU · xfast (Apple M4 Pro) · 14 GB 6m 53s $0.5507
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
Each row's best-value machine of this size for that build (lowest time × cost, ranked on p90), in that order, named under the provider. Time is the median whole job, cost is that job at the provider's per-minute price (RunsOn at EC2 spot). Yellow marks the fastest and the cheapest. Apple silicon and tmpfs (workspace in RAM) machines are rows of their own; a faded row is provisional or a machine of another size, and takes no place. A row that did not finish a build stays listed with the reason.
Full build benchmark every phase, p90 and worst runs, queue under a burst, run-to-run spread, bigger runners
Real buildstime and cost per buildlower is betterarm64 · 8 vCPU Share Save image
Rust cold build of spotify-player: cargo test and clippy
RunsOn tmpfs is the best value and the cheapest; Avrea is the fastest.
- RunsOntmpfs m9g.2xlarge · 32 GB 1m 27s $0.0037
- RunsOn m9g.2xlarge · EBS gp3 · 32 GB 1m 28s $0.0037
- Warpbuild 32 GB 2m 11s $0.0262
- Blacksmith 32 GB 2m 51s $0.0284
- Ubicloud standard · 32 GB 3m 51s $0.0192
- Avrea 32 GB 1m 07s $0.0532
- Namespace 8x32 · 32 GB 4m 05s $0.0980
- WarpbuildApple silicon 12 vCPU · xfast (Apple M4 Pro) · 28 GB 1m 10s $0.1867
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- GitHub no 8 vCPU arm64 runner
- NamespaceApple silicon no 8 vCPU arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
TypeScript cold bun install and typecheck of opencode
RunsOn tmpfs is the best value and the cheapest; Avrea is the fastest.
- RunsOntmpfs m9g.2xlarge · 32 GB 46s $0.0022
- RunsOn m9gd.2xlarge · local NVMe · 32 GB 44s $0.0032
- Warpbuild 32 GB 1m 04s $0.0128
- Ubicloud standard · 32 GB 1m 43s $0.0086
- Blacksmith 32 GB 1m 24s $0.0140
- Namespace 8x16 · 16 GB 2m 03s $0.0246
- Avrea 32 GB 39s $0.0312
- WarpbuildApple silicon 12 vCPU · xfast (Apple M4 Pro) · 28 GB 41s $0.1080
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- GitHub no 8 vCPU arm64 runner
- NamespaceApple silicon no 8 vCPU arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
Docker PostHog's production image, built cold with buildx
RunsOn tmpfs is the best value and the cheapest; Avrea is the fastest.
- RunsOntmpfs m9g.2xlarge · 32 GB 4m 19s $0.0101
- RunsOn m9gd.2xlarge · local NVMe · 32 GB 4m 41s $0.0163
- Warpbuild 32 GB 6m 57s $0.0834
- Blacksmith 32 GB 7m 43s $0.0772
- Ubicloud standard · 32 GB 11m 38s $0.0582
- Avrea 32 GB 3m 59s $0.1912
- Namespace 8x16 · 16 GB 11m 02s $0.1323
- WarpbuildApple silicon 12 vCPU · xfast (Apple M4 Pro) · 28 GB 5m 22s $0.8587
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- GitHub no 8 vCPU arm64 runner
- NamespaceApple silicon no 8 vCPU arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
Each row's best-value machine of this size for that build (lowest time × cost, ranked on p90), in that order, named under the provider. Time is the median whole job, cost is that job at the provider's per-minute price (RunsOn at EC2 spot). Yellow marks the fastest and the cheapest. Apple silicon and tmpfs (workspace in RAM) machines are rows of their own; a faded row is provisional or a machine of another size, and takes no place. A row that did not finish a build stays listed with the reason.
Full build benchmark every phase, p90 and worst runs, queue under a burst, run-to-run spread, bigger runners
CPU
CPUsingle-thread PassMarkhigher is betterx64 · 2 vCPU Share Save image
Namespace and Avrea tie for the fastest core.
RunsOn is 3rd of 9, 7% below the leader. Namespace to GitHub: 2.0× apart.
- Namespace 2x8 · 8 GB 4,543
- Avrea 8 GB 4,358
- RunsOn m8azn.large · EBS gp3 · 8 GB 4,234
- Warpbuild 8 GB 3,936
- RunsOntmpfs r8a.large · 16 GB 3,797
- Blacksmith 8 GB 3,725
- Ubicloud standard · 8 GB 2,866
- AWS CodeBuild small · 3 GB 2,602
- boat 4 GB 2,302
- GitHub 8 GB 2,243
- Sengi premium · 8 GB 3,705 provisional
- NamespaceApple silicon arm64 only
- StarSling no 2 vCPU x64 runner
- WarpbuildApple silicon arm64 only
The figure is the score a runner's jobs reach 9 times out of 10 (p10), so a label that sometimes lands on an older CPU ranks below one that never does. Over the bar: the lowest and the highest job (the line), 9 in 10 to 1 in 10 (the box) and the median (the tick). A long box or line is a CPU lottery. The trend is the daily median, on a scale of at least ±10%, so only a real change shows as a slope.
CPUsingle-thread PassMarkhigher is betterx64 · 4 vCPU Share Save image
RunsOn, Namespace and Avrea tie for the fastest core.
Namespace to GitHub: 1.9× apart.
- Namespace 4x16 · 16 GB 4,441
- Avrea 16 GB 4,401
- RunsOn m8azn.xlarge · gp3 1000 MiB/s · 16 GB 4,255
- Warpbuild 16 GB 4,081
- RunsOntmpfs r8a.xlarge · 32 GB 3,802
- Blacksmith 16 GB 3,614
- StarSling 16 GB 3,295
- Ubicloud premium · 16 GB 3,146
- boat 8 GB 3,015
- AWS CodeBuild medium · 8 GB 2,621
- GitHub 16 GB 2,365
- Sengi premium · 16 GB too few days 3,724 provisional
- NamespaceApple silicon arm64 only
- WarpbuildApple silicon arm64 only
The figure is the score a runner's jobs reach 9 times out of 10 (p10), so a label that sometimes lands on an older CPU ranks below one that never does. Over the bar: the lowest and the highest job (the line), 9 in 10 to 1 in 10 (the box) and the median (the tick). A long box or line is a CPU lottery. The trend is the daily median, on a scale of at least ±10%, so only a real change shows as a slope.
CPUsingle-thread PassMarkhigher is betterx64 · 8 vCPU Share Save image
Namespace and Avrea tie for the fastest core.
RunsOn is 4th of 10, 17% below the leader. Namespace to GitHub: 1.9× apart.
- Namespace 8x32 · 32 GB 4,562
- Avrea 32 GB 4,434
- Warpbuild 32 GB 4,040
- RunsOntmpfs r8a.2xlarge · 64 GB 3,798
- RunsOn m8a.2xlarge · EBS gp3 · 32 GB 3,740
- Blacksmith 32 GB 3,665
- Ubicloud premium · 32 GB 3,616
- StarSling 32 GB 3,412
- boat 16 GB 2,880
- AWS CodeBuild large · 16 GB 2,626
- GitHub 32 GB 2,350
- Sengi premium · 32 GB too few days 3,745 provisional
- NamespaceApple silicon arm64 only
- WarpbuildApple silicon arm64 only
The figure is the score a runner's jobs reach 9 times out of 10 (p10), so a label that sometimes lands on an older CPU ranks below one that never does. Over the bar: the lowest and the highest job (the line), 9 in 10 to 1 in 10 (the box) and the median (the tick). A long box or line is a CPU lottery. The trend is the daily median, on a scale of at least ±10%, so only a real change shows as a slope.
CPUsingle-thread PassMarkhigher is betterarm64 · 2 vCPU Share Save image
Avrea and Namespace Apple silicon tie for the fastest core.
RunsOn is 3rd of 7, 48% below the leader. Avrea to Namespace: 4.1× apart.
- Avrea 8 GB 4,744
- NamespaceApple silicon 2x8 · 8 GB 4,594
- RunsOn m9gd.large · local NVMe · 8 GB 2,475
- RunsOntmpfs r9g.large · 16 GB 2,466
- Warpbuild 8 GB 1,939
- GitHub 8 GB 1,869
- Blacksmith 6 GB 1,330
- Ubicloud standard · 8 GB 1,297
- Namespace 2x8 · 8 GB 1,146
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
- WarpbuildApple silicon no 2 vCPU arm64 runner
The figure is the score a runner's jobs reach 9 times out of 10 (p10), so a label that sometimes lands on an older CPU ranks below one that never does. Over the bar: the lowest and the highest job (the line), 9 in 10 to 1 in 10 (the box) and the median (the tick). A long box or line is a CPU lottery. The trend is the daily median, on a scale of at least ±10%, so only a real change shows as a slope.
CPUsingle-thread PassMarkhigher is betterarm64 · 4 vCPU Share Save image
Avrea and Namespace Apple silicon tie for the fastest core.
RunsOn is 3rd of 7, 49% below the leader. Avrea to Namespace: 4.1× apart.
- Avrea 16 GB 4,816
- NamespaceApple silicon 4x16 · 16 GB 4,584
- RunsOn m9gd.xlarge · local NVMe · 16 GB 2,472
- RunsOntmpfs r9g.xlarge · 32 GB 2,467
- Warpbuild 16 GB 1,940
- GitHub 16 GB 1,870
- Blacksmith 16 GB 1,340
- Ubicloud standard · 16 GB 1,286
- Namespace 4x16 · 16 GB 1,162
- WarpbuildApple silicon 6 vCPU · xfast (Apple M4 Pro) · 14 GB 3,006 not ranked
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
The figure is the score a runner's jobs reach 9 times out of 10 (p10), so a label that sometimes lands on an older CPU ranks below one that never does. Over the bar: the lowest and the highest job (the line), 9 in 10 to 1 in 10 (the box) and the median (the tick). A long box or line is a CPU lottery. The trend is the daily median, on a scale of at least ±10%, so only a real change shows as a slope.
CPUsingle-thread PassMarkhigher is betterarm64 · 8 vCPU Share Save image
Avrea has the fastest core.
RunsOn is 2nd of 6, 49% below the leader. Avrea to Namespace: 4.1× apart.
- Avrea 32 GB 4,815
- RunsOn m9g.2xlarge · EBS gp3 · 32 GB 2,472
- RunsOntmpfs m9g.2xlarge · 32 GB 2,471
- Warpbuild 32 GB 1,940
- Blacksmith 32 GB 1,344
- Ubicloud standard · 32 GB 1,298
- Namespace 8x16 · 16 GB 1,166
- WarpbuildApple silicon 12 vCPU · xfast (Apple M4 Pro) · 28 GB 3,145 not ranked
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- GitHub no 8 vCPU arm64 runner
- NamespaceApple silicon no 8 vCPU arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
The figure is the score a runner's jobs reach 9 times out of 10 (p10), so a label that sometimes lands on an older CPU ranks below one that never does. Over the bar: the lowest and the highest job (the line), 9 in 10 to 1 in 10 (the box) and the median (the tick). A long box or line is a CPU lottery. The trend is the daily median, on a scale of at least ±10%, so only a real change shows as a slope.
Price
Priceper minutelower is betterx64 · 2 vCPU Share Save image
boat has the lowest price per minute.
RunsOn is 2nd of 10, the leader is 2.5× lower. boat to Namespace: 20× apart.
- boat 4 GB $0.0003
- RunsOn m8a.large · EBS gp3 · 8 GB $0.0008 $0.0023 on-demand
- RunsOntmpfs r8a.large · 16 GB $0.0008 $0.0030 on-demand
- Sengi basic · 8 GB $0.0010
- Ubicloud standard · 8 GB $0.0013
- Avrea 8 GB $0.0040
- Blacksmith 8 GB $0.0040
- Warpbuild 8 GB $0.0040
- AWS CodeBuild small · 3 GB $0.0050
- GitHub 8 GB $0.0060
- Namespace 2x8 · 8 GB $0.0060
- NamespaceApple silicon arm64 only
- StarSling no 2 vCPU x64 runner
- WarpbuildApple silicon arm64 only
Pay-as-you-go price of a 2 vCPU runner. RunsOn is the EC2 + EBS bill in your own AWS account, at spot (the bar) and on-demand (the outline); its license is flat, like the other providers' plan fees.
Priceper minutelower is betterx64 · 4 vCPU Share Save image
boat has the lowest price per minute.
RunsOn is 2nd of 11, 94% above the leader. boat to Namespace: 20× apart.
- boat 8 GB $0.0006
- RunsOn m8i-flex.xlarge · EBS gp3 · 16 GB $0.0012 $0.0037 on-demand
- RunsOntmpfs r8a.xlarge · 32 GB $0.0015 $0.0056 on-demand
- Sengi basic · 16 GB $0.0020
- Ubicloud standard · 16 GB $0.0025
- Avrea 16 GB $0.0080
- Blacksmith 16 GB $0.0080
- StarSling 16 GB $0.0080
- Warpbuild 16 GB $0.0080
- AWS CodeBuild medium · 8 GB $0.0100
- GitHub 16 GB $0.0120
- Namespace 4x16 · 16 GB $0.0120
- NamespaceApple silicon arm64 only
- WarpbuildApple silicon arm64 only
Pay-as-you-go price of a 4 vCPU runner. RunsOn is the EC2 + EBS bill in your own AWS account, at spot (the bar) and on-demand (the outline); its license is flat, like the other providers' plan fees.
Priceper minutelower is betterx64 · 8 vCPU Share Save image
boat has the lowest price per minute.
RunsOn is 2nd of 11, 98% above the leader. boat to Namespace: 20× apart.
- boat 16 GB $0.0012
- RunsOn m8a.2xlarge · EBS gp3 · 32 GB $0.0024 $0.0084 on-demand
- RunsOntmpfs m8a.2xlarge · 32 GB $0.0024 $0.0084 on-demand
- Sengi basic · 32 GB $0.0040
- Ubicloud standard · 32 GB $0.0050
- Avrea 32 GB $0.0160
- Blacksmith 32 GB $0.0160
- StarSling 32 GB $0.0160
- Warpbuild 32 GB $0.0160
- AWS CodeBuild large · 16 GB $0.0200
- GitHub 32 GB $0.0220
- Namespace 8x32 · 32 GB $0.0240
- NamespaceApple silicon arm64 only
- WarpbuildApple silicon arm64 only
Pay-as-you-go price of a 8 vCPU runner. RunsOn is the EC2 + EBS bill in your own AWS account, at spot (the bar) and on-demand (the outline); its license is flat, like the other providers' plan fees.
Priceper minutelower is betterarm64 · 2 vCPU Share Save image
RunsOn has the lowest price per minute.
RunsOn to Namespace Apple silicon: 52× apart.
- RunsOn m9gd.large · local NVMe · 8 GB $0.0008 $0.0024 on-demand
- RunsOntmpfs r9g.large · 16 GB $0.0010 $0.0024 on-demand
- Ubicloud standard · 8 GB $0.0013
- Blacksmith 6 GB $0.0025
- Warpbuild 8 GB $0.0030
- GitHub 8 GB $0.0050
- Namespace 2x8 · 8 GB $0.0060
- Avrea 8 GB $0.0120
- NamespaceApple silicon 2x8 · 8 GB $0.0420
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
- WarpbuildApple silicon no 2 vCPU arm64 runner
Pay-as-you-go price of a 2 vCPU runner. RunsOn is the EC2 + EBS bill in your own AWS account, at spot (the bar) and on-demand (the outline); its license is flat, like the other providers' plan fees.
Priceper minutelower is betterarm64 · 4 vCPU Share Save image
RunsOn has the lowest price per minute.
RunsOn to Namespace Apple silicon: 64× apart.
- RunsOn m9gd.xlarge · local NVMe · 16 GB $0.0013 $0.0045 on-demand
- RunsOntmpfs m9g.xlarge · 16 GB $0.0017 $0.0036 on-demand
- Ubicloud standard · 16 GB $0.0025
- Blacksmith 16 GB $0.0050
- Warpbuild 16 GB $0.0060
- GitHub 16 GB $0.0080
- Namespace 4x16 · 16 GB $0.0120
- Avrea 16 GB $0.0240
- NamespaceApple silicon 4x16 · 16 GB $0.0840
- WarpbuildApple silicon 6 vCPU · xfast (Apple M4 Pro) · 14 GB $0.0800 not ranked
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
Pay-as-you-go price of a 4 vCPU runner. RunsOn is the EC2 + EBS bill in your own AWS account, at spot (the bar) and on-demand (the outline); its license is flat, like the other providers' plan fees.
Priceper minutelower is betterarm64 · 8 vCPU Share Save image
RunsOn and RunsOn tmpfs tie for the lowest price per minute.
RunsOn to Avrea: 21× apart.
- RunsOn m9g.2xlarge · EBS gp3 · 32 GB $0.0022 $0.0068 on-demand
- RunsOntmpfs m9g.2xlarge · 32 GB $0.0022 $0.0068 on-demand
- Ubicloud standard · 32 GB $0.0050
- Blacksmith 32 GB $0.0100
- Warpbuild 32 GB $0.0120
- Namespace 8x32 · 32 GB $0.0240
- Avrea 32 GB $0.0480
- WarpbuildApple silicon 12 vCPU · xfast (Apple M4 Pro) · 28 GB $0.1600 not ranked
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- GitHub no 8 vCPU arm64 runner
- NamespaceApple silicon no 8 vCPU arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
Pay-as-you-go price of a 8 vCPU runner. RunsOn is the EC2 + EBS bill in your own AWS account, at spot (the bar) and on-demand (the outline); its license is flat, like the other providers' plan fees.
Network
Networkdownload from GitHubhigher is betterx64 · 2 vCPU Share Save image
RunsOn tmpfs downloads from GitHub fastest.
RunsOn tmpfs to Ubicloud: 6.1× apart.
- RunsOntmpfs r8a.large · 16 GB 303 MiB/s
- RunsOn median of 9 machines 205 MiB/s
- Namespace 2x8 · 8 GB 163 MiB/s
- AWS CodeBuild small · 3 GB 155 MiB/s
- Warpbuild 8 GB 111 MiB/s
- Avrea 8 GB 97 MiB/s
- Blacksmith 8 GB 95 MiB/s
- boat 4 GB 72 MiB/s
- GitHub 8 GB 72 MiB/s
- Ubicloud median of 2 machines 50 MiB/s
- Sengi median of 2 machines 51 MiB/s provisional
- NamespaceApple silicon arm64 only
- StarSling no 2 vCPU x64 runner
- WarpbuildApple silicon arm64 only
A large file fetched from GitHub inside the job. Downloads vary from machine to machine, so a bar is the median of the line's machines of this size and the mark over it runs from its slowest to its fastest.
Networkdownload from GitHubhigher is betterx64 · 4 vCPU Share Save image
StarSling and AWS CodeBuild tie for the fastest download from GitHub.
RunsOn is 3rd of 10, 14% below the leader. StarSling to Ubicloud: 4.0× apart.
- StarSling 16 GB 217 MiB/s
- AWS CodeBuild medium · 8 GB 213 MiB/s
- RunsOn median of 8 machines 186 MiB/s
- Avrea 16 GB 163 MiB/s
- RunsOntmpfs r8a.xlarge · 32 GB 159 MiB/s
- Namespace median of 2 machines 124 MiB/s
- GitHub 16 GB 100 MiB/s
- Warpbuild 16 GB 99 MiB/s
- Blacksmith 16 GB 83 MiB/s
- boat 8 GB 61 MiB/s
- Ubicloud median of 2 machines 54 MiB/s
- Sengi median of 2 machines 82 MiB/s provisional
- NamespaceApple silicon arm64 only
- WarpbuildApple silicon arm64 only
A large file fetched from GitHub inside the job. Downloads vary from machine to machine, so a bar is the median of the line's machines of this size and the mark over it runs from its slowest to its fastest.
Networkdownload from GitHubhigher is betterx64 · 8 vCPU Share Save image
RunsOn and RunsOn tmpfs tie for the fastest download from GitHub.
RunsOn to Ubicloud: 3.6× apart.
- RunsOn median of 7 machines 178 MiB/s
- RunsOntmpfs median of 2 machines 172 MiB/s
- AWS CodeBuild large · 16 GB 161 MiB/s
- Avrea 32 GB 150 MiB/s
- Namespace median of 2 machines 137 MiB/s
- StarSling 32 GB 130 MiB/s
- GitHub 32 GB 107 MiB/s
- Blacksmith 32 GB 105 MiB/s
- Warpbuild 32 GB 101 MiB/s
- boat 16 GB 79 MiB/s
- Ubicloud median of 2 machines 50 MiB/s
- Sengi median of 2 machines 56 MiB/s provisional
- NamespaceApple silicon arm64 only
- WarpbuildApple silicon arm64 only
A large file fetched from GitHub inside the job. Downloads vary from machine to machine, so a bar is the median of the line's machines of this size and the mark over it runs from its slowest to its fastest.
Networkdownload from GitHubhigher is betterarm64 · 2 vCPU Share Save image
RunsOn tmpfs downloads from GitHub fastest.
RunsOn tmpfs to Ubicloud: 3.3× apart.
- RunsOntmpfs r9g.large · 16 GB 191 MiB/s
- RunsOn median of 3 machines 177 MiB/s
- Warpbuild 8 GB 167 MiB/s
- NamespaceApple silicon 2x8 · 8 GB 143 MiB/s
- Avrea 8 GB 141 MiB/s
- Namespace 2x8 · 8 GB 122 MiB/s
- GitHub 8 GB 120 MiB/s
- Blacksmith 6 GB 92 MiB/s
- Ubicloud standard · 8 GB 58 MiB/s
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
- WarpbuildApple silicon no 2 vCPU arm64 runner
A large file fetched from GitHub inside the job. Downloads vary from machine to machine, so a bar is the median of the line's machines of this size and the mark over it runs from its slowest to its fastest.
Networkdownload from GitHubhigher is betterarm64 · 4 vCPU Share Save image
RunsOn and RunsOn tmpfs tie for the fastest download from GitHub.
RunsOn to Ubicloud: 2.3× apart.
- RunsOn median of 3 machines 172 MiB/s
- RunsOntmpfs median of 2 machines 167 MiB/s
- Warpbuild 16 GB 158 MiB/s
- NamespaceApple silicon 4x16 · 16 GB 127 MiB/s
- Namespace median of 2 machines 125 MiB/s
- GitHub 16 GB 118 MiB/s
- Blacksmith 16 GB 97 MiB/s
- Avrea 16 GB 86 MiB/s
- Ubicloud standard · 16 GB 74 MiB/s
- WarpbuildApple silicon 6 vCPU · xfast (Apple M4 Pro) · 14 GB 68 MiB/s not ranked
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
A large file fetched from GitHub inside the job. Downloads vary from machine to machine, so a bar is the median of the line's machines of this size and the mark over it runs from its slowest to its fastest.
Networkdownload from GitHubhigher is betterarm64 · 8 vCPU Share Save image
RunsOn tmpfs downloads from GitHub fastest.
RunsOn tmpfs to Ubicloud: 4.4× apart.
- RunsOntmpfs m9g.2xlarge · 32 GB 253 MiB/s
- Warpbuild 32 GB 198 MiB/s
- RunsOn median of 3 machines 161 MiB/s
- Namespace median of 2 machines 134 MiB/s
- Blacksmith 32 GB 101 MiB/s
- Avrea 32 GB 84 MiB/s
- Ubicloud standard · 32 GB 57 MiB/s
- WarpbuildApple silicon 12 vCPU · xfast (Apple M4 Pro) · 28 GB 70 MiB/s not ranked
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- GitHub no 8 vCPU arm64 runner
- NamespaceApple silicon no 8 vCPU arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
A large file fetched from GitHub inside the job. Downloads vary from machine to machine, so a bar is the median of the line's machines of this size and the mark over it runs from its slowest to its fastest.
Disk
Diskrandom reads per secondhigher is betterx64 · 2 vCPU Share Save image
Namespace has the fastest disk on small reads.
RunsOn is 5th of 9, the leader is 8.4× higher. Namespace to AWS CodeBuild: 60× apart.
- Namespace 2x8 · 8 GB 105,054
- Blacksmith 8 GB 71,426
- Avrea 8 GB 53,598
- Warpbuild 8 GB 24,481
- RunsOn i7i.large · local NVMe · 16 GB 12,510
- boat 4 GB 10,783
- Ubicloud standard · 8 GB 6,889
- GitHub 8 GB 6,868
- AWS CodeBuild small · 3 GB 1,763
- Sengi premium · 8 GB 13,893 provisional
- NamespaceApple silicon arm64 only
- RunsOntmpfs workspace in RAM, no disk
- StarSling no 2 vCPU x64 runner
- WarpbuildApple silicon arm64 only
4k random reads one at a time (fio, QD1) in the job's workspace, each provider's best disk. It is a disk's limit: a build often never reaches it, so read it next to the builds. A tmpfs workspace is RAM and has no disk figure.
Diskrandom reads per secondhigher is betterx64 · 4 vCPU Share Save image
Namespace has the fastest disk on small reads.
RunsOn is 6th of 10, the leader is 8.8× higher. Namespace to AWS CodeBuild: 58× apart.
- Namespace 4x8 · 8 GB 104,646
- Blacksmith 16 GB 82,000
- Warpbuild 16 GB 63,558
- Avrea 16 GB 49,550
- StarSling 16 GB 32,038
- RunsOn i7i.xlarge · local NVMe · 32 GB 11,950
- boat 8 GB 10,338
- Ubicloud premium · 16 GB 7,801
- GitHub 16 GB 7,332
- AWS CodeBuild medium · 8 GB 1,794
- Sengi premium · 16 GB 13,632 provisional
- NamespaceApple silicon arm64 only
- RunsOntmpfs workspace in RAM, no disk
- WarpbuildApple silicon arm64 only
4k random reads one at a time (fio, QD1) in the job's workspace, each provider's best disk. It is a disk's limit: a build often never reaches it, so read it next to the builds. A tmpfs workspace is RAM and has no disk figure.
Diskrandom reads per secondhigher is betterx64 · 8 vCPU Share Save image
Namespace has the fastest disk on small reads.
RunsOn is 7th of 10, the leader is 9.6× higher. Namespace to AWS CodeBuild: 60× apart.
- Namespace 8x32 · 32 GB 106,228
- Blacksmith 32 GB 74,095
- Avrea 32 GB 52,121
- StarSling 32 GB 52,058
- Warpbuild 32 GB 37,153
- boat 16 GB 11,383
- RunsOn m8id.2xlarge · local NVMe · 32 GB 11,083
- Ubicloud premium · 32 GB 9,543
- GitHub 32 GB 7,559
- AWS CodeBuild large · 16 GB 1,758
- Sengi premium · 32 GB 13,912 provisional
- NamespaceApple silicon arm64 only
- RunsOntmpfs workspace in RAM, no disk
- WarpbuildApple silicon arm64 only
4k random reads one at a time (fio, QD1) in the job's workspace, each provider's best disk. It is a disk's limit: a build often never reaches it, so read it next to the builds. A tmpfs workspace is RAM and has no disk figure.
Diskrandom reads per secondhigher is betterarm64 · 2 vCPU Share Save image
Blacksmith has the fastest disk on small reads.
RunsOn is 4th of 7, the leader is 2.3× higher. Blacksmith to Warpbuild: 17× apart.
- Blacksmith 6 GB 31,028
- Avrea 8 GB 20,909
- NamespaceApple silicon 2x8 · 8 GB 19,829
- Namespace 2x8 · 8 GB 15,941
- RunsOn m9gd.large · local NVMe · 8 GB 13,533
- Ubicloud standard · 8 GB 8,452
- GitHub 8 GB 8,119
- Warpbuild 8 GB 1,779
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- RunsOntmpfs workspace in RAM, no disk
- Sengi no arm64 runner
- StarSling no arm64 runner
- WarpbuildApple silicon no 2 vCPU arm64 runner
4k random reads one at a time (fio, QD1) in the job's workspace, each provider's best disk. It is a disk's limit: a build often never reaches it, so read it next to the builds. A tmpfs workspace is RAM and has no disk figure.
Diskrandom reads per secondhigher is betterarm64 · 4 vCPU Share Save image
Blacksmith has the fastest disk on small reads.
RunsOn is 4th of 7, the leader is 2.4× higher. Blacksmith to Warpbuild: 18× apart.
- Blacksmith 16 GB 32,691
- Avrea 16 GB 20,301
- NamespaceApple silicon 4x16 · 16 GB 18,570
- Namespace 4x16 · 16 GB 16,759
- RunsOn m9gd.xlarge · local NVMe · 16 GB 13,368
- Ubicloud standard · 16 GB 8,334
- GitHub 16 GB 7,694
- Warpbuild 16 GB 1,812
- WarpbuildApple silicon 6 vCPU · xfast (Apple M4 Pro) · 14 GB 29,080 not ranked
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- RunsOntmpfs workspace in RAM, no disk
- Sengi no arm64 runner
- StarSling no arm64 runner
4k random reads one at a time (fio, QD1) in the job's workspace, each provider's best disk. It is a disk's limit: a build often never reaches it, so read it next to the builds. A tmpfs workspace is RAM and has no disk figure.
Diskrandom reads per secondhigher is betterarm64 · 8 vCPU Share Save image
Blacksmith has the fastest disk on small reads.
RunsOn is 4th of 6, the leader is 2.4× higher. Blacksmith to Warpbuild: 18× apart.
- Blacksmith 32 GB 31,905
- Avrea 32 GB 20,397
- Namespace 8x16 · 16 GB 17,452
- RunsOn m9gd.2xlarge · local NVMe · 32 GB 13,462
- Ubicloud standard · 32 GB 8,453
- Warpbuild 32 GB 1,781
- WarpbuildApple silicon 12 vCPU · xfast (Apple M4 Pro) · 28 GB 30,102 not ranked
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- GitHub no 8 vCPU arm64 runner
- NamespaceApple silicon no 8 vCPU arm64 runner
- RunsOntmpfs workspace in RAM, no disk
- Sengi no arm64 runner
- StarSling no arm64 runner
4k random reads one at a time (fio, QD1) in the job's workspace, each provider's best disk. It is a disk's limit: a build often never reaches it, so read it next to the builds. A tmpfs workspace is RAM and has no disk figure.
Cache
Cacherestoring 4 GBlower is betterx64 · 2 vCPU Share Save image
RunsOn tmpfs restores a 4 GB cache fastest.
RunsOn tmpfs to boat: 13× apart.
- RunsOntmpfs r8a.large · actions/cache 7s
- Blacksmith actions/cache 9s
- Avrea actions/cache 12s
- RunsOn r8a.large · EBS gp3 · actions/cache 17s
- Namespace 2x8 · actions/cache 21s
- Warpbuild actions/cache 26s
- AWS CodeBuild small · actions/cache 52s
- Ubicloud standard · actions/cache 59s
- GitHub 2-core · actions/cache 1m 00s
- boat actions/cache 1m 32s
- NamespaceApple silicon arm64 only
- Sengi no cache result
- StarSling no 2 vCPU x64 runner
- WarpbuildApple silicon arm64 only
Restore of a 4 GB cache: each provider's fastest cache action on its standard storage, named under the provider; the tmpfs row restores into RAM.
Cacherestoring 4 GBlower is betterx64 · 4 vCPU Share Save image
RunsOn tmpfs restores a 4 GB cache fastest.
RunsOn tmpfs to boat: 19× apart.
- RunsOntmpfs r8a.xlarge · actions/cache 5s
- Avrea actions/cache 10s
- Blacksmith actions/cache 10s
- RunsOn m8a.xlarge · EBS gp3 · actions/cache 16s
- Namespace 4x8 · actions/cache 18s
- Warpbuild actions/cache 21s
- StarSling actions/cache 23s
- AWS CodeBuild medium · actions/cache 35s
- GitHub 4-core · actions/cache 38s
- Ubicloud premium · actions/cache 56s
- boat actions/cache 1m 36s
- Sengi premium · actions/cache 7s provisional
- NamespaceApple silicon arm64 only
- WarpbuildApple silicon arm64 only
Restore of a 4 GB cache: each provider's fastest cache action on its standard storage, named under the provider; the tmpfs row restores into RAM.
Cacherestoring 4 GBlower is betterx64 · 8 vCPU Share Save image
RunsOn tmpfs restores a 4 GB cache fastest.
RunsOn tmpfs to boat: 16× apart.
- RunsOntmpfs m8a.2xlarge · actions/cache 5s
- Blacksmith actions/cache 8s
- Avrea actions/cache 10s
- RunsOn m8a.2xlarge · EBS gp3 · actions/cache 11s
- Namespace 8x16 · actions/cache 18s
- Warpbuild actions/cache 23s
- StarSling actions/cache 24s
- GitHub 8-core · actions/cache 33s
- AWS CodeBuild large · actions/cache 35s
- Ubicloud standard · actions/cache 55s
- boat actions/cache 1m 11s
- Sengi premium · actions/cache 6s provisional
- NamespaceApple silicon arm64 only
- WarpbuildApple silicon arm64 only
Restore of a 4 GB cache: each provider's fastest cache action on its standard storage, named under the provider; the tmpfs row restores into RAM.
Cacherestoring 4 GBlower is betterarm64 · 2 vCPU Share Save image
RunsOn tmpfs restores a 4 GB cache fastest.
RunsOn tmpfs to Ubicloud: 8.0× apart.
- RunsOntmpfs r9g.large · actions/cache 8s
- Avrea actions/cache 12s
- RunsOn m9g.large · EBS gp3 · actions/cache 19s
- Blacksmith actions/cache 22s
- NamespaceApple silicon 2x8 · actions/cache 24s
- Warpbuild actions/cache 28s
- Namespace 2x8 · actions/cache 43s
- GitHub 2-core · actions/cache 53s
- Ubicloud standard · actions/cache 1m 04s
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
- WarpbuildApple silicon no 2 vCPU arm64 runner
Restore of a 4 GB cache: each provider's fastest cache action on its standard storage, named under the provider; the tmpfs row restores into RAM.
Cacherestoring 4 GBlower is betterarm64 · 4 vCPU Share Save image
RunsOn tmpfs restores a 4 GB cache fastest.
RunsOn tmpfs to Ubicloud: 12× apart.
- RunsOntmpfs r9g.xlarge · actions/cache 5s
- Avrea actions/cache 10s
- RunsOn m9g.xlarge · EBS gp3 · actions/cache 16s
- Blacksmith actions/cache 20s
- NamespaceApple silicon 4x16 · actions/cache 21s
- Warpbuild actions/cache 27s
- GitHub 4-core · actions/cache 37s
- Namespace 4x8 · actions/cache 39s
- Ubicloud standard · actions/cache 1m 01s
- WarpbuildApple silicon 6 vCPU · xfast (Apple M4 Pro) · actions/cache 37s not ranked
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
Restore of a 4 GB cache: each provider's fastest cache action on its standard storage, named under the provider; the tmpfs row restores into RAM.
Cacherestoring 4 GBlower is betterarm64 · 8 vCPU Share Save image
RunsOn tmpfs restores a 4 GB cache fastest.
RunsOn tmpfs to Ubicloud: 13× apart.
- RunsOntmpfs m9g.2xlarge · actions/cache 4s
- Avrea actions/cache 10s
- RunsOn m9g.2xlarge · EBS gp3 · actions/cache 10s
- Blacksmith actions/cache 17s
- Warpbuild actions/cache 21s
- Namespace 8x16 · actions/cache 34s
- Ubicloud standard · actions/cache 50s
- WarpbuildApple silicon 12 vCPU · xfast (Apple M4 Pro) · WarpBuilds/cache 1m 06s not ranked
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- GitHub no 8 vCPU arm64 runner
- NamespaceApple silicon no 8 vCPU arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
Restore of a 4 GB cache: each provider's fastest cache action on its standard storage, named under the provider; the tmpfs row restores into RAM.
Predictability
Predictabilityrun to runlower is betterx64 · 2 vCPU Share Save image
Avrea has the steadiest build times.
RunsOn is 2nd of 6, at +9%.
- Avrea 8 GB · worst run 1.1× +3%
- RunsOntmpfs r8a.large · 16 GB · worst run 1.1× +9%
- RunsOn 2 machines · worst run 1.2× +9%
- GitHub 8 GB · worst run 1.3× +14%
- Warpbuild 8 GB · worst run 1.3× +18%
- Ubicloud standard · 8 GB · worst run 1.4× +20%
- Blacksmith 8 GB · worst run 1.7× +31%
- AWS CodeBuild small · 3 GB · worst run 1.1× +8% provisional
- boat 4 GB · worst run 1.2× +17% provisional
- Namespace 2x8 · 8 GB · worst run 1.3× +25% provisional
- NamespaceApple silicon arm64 only
- Sengi under 3 runs of any build
- StarSling no 2 vCPU x64 runner
- WarpbuildApple silicon arm64 only
How much a build's time moves from run to run on the same machine: the run that 9 runs in 10 beat (p90), over the typical run (p50), for each build on the machine Real builds picks, averaged over the builds with 3 runs or more. +10% means one run in ten took more than 10% longer than usual. A line measured on one build is shown, not ranked.
Predictabilityrun to runlower is betterx64 · 4 vCPU Share Save image
StarSling has the steadiest build times.
RunsOn is 2nd of 9, at +3%.
- StarSling 16 GB · worst run 1.0× +2%
- RunsOntmpfs r8a.xlarge · 32 GB · worst run 1.0× +3%
- RunsOn 2 machines · worst run 1.1× +3%
- Avrea 16 GB · worst run 1.1× +3%
- Warpbuild 16 GB · worst run 1.3× +8%
- Namespace 2 machines · worst run 1.4× +12%
- GitHub 16 GB · worst run 1.2× +14%
- Ubicloud standard · 16 GB · worst run 1.3× +16%
- Blacksmith 16 GB · worst run 2.8× +44%
- boat 8 GB · worst run 3.6× +117%
- AWS CodeBuild medium · 8 GB · worst run 1.1× +6% provisional
- NamespaceApple silicon arm64 only
- Sengi under 3 runs of any build
- WarpbuildApple silicon arm64 only
How much a build's time moves from run to run on the same machine: the run that 9 runs in 10 beat (p90), over the typical run (p50), for each build on the machine Real builds picks, averaged over the builds with 3 runs or more. +10% means one run in ten took more than 10% longer than usual. A line measured on one build is shown, not ranked.
Predictabilityrun to runlower is betterx64 · 8 vCPU Share Save image
Avrea has the steadiest build times.
RunsOn is 3rd of 10, at +5%.
- Avrea 32 GB · worst run 1.1× +2%
- AWS CodeBuild large · 16 GB · worst run 1.1× +3%
- RunsOntmpfs m8a.2xlarge · 32 GB · worst run 1.2× +5%
- Warpbuild 32 GB · worst run 1.2× +5%
- RunsOn 2 machines · worst run 1.3× +6%
- StarSling 32 GB · worst run 1.2× +11%
- GitHub 32 GB · worst run 1.2× +13%
- Namespace 2 machines · worst run 1.6× +14%
- Ubicloud premium · 32 GB · worst run 1.4× +16%
- Blacksmith 32 GB · worst run 1.9× +31%
- boat 16 GB · worst run 6.1× +197%
- NamespaceApple silicon arm64 only
- Sengi under 3 runs of any build
- WarpbuildApple silicon arm64 only
How much a build's time moves from run to run on the same machine: the run that 9 runs in 10 beat (p90), over the typical run (p50), for each build on the machine Real builds picks, averaged over the builds with 3 runs or more. +10% means one run in ten took more than 10% longer than usual. A line measured on one build is shown, not ranked.
Predictabilityrun to runlower is betterarm64 · 2 vCPU Share Save image
RunsOn and Blacksmith tie for the steadiest build times.
- RunsOn m9gd.large · local NVMe · 8 GB · worst run 1.1× +5%
- Blacksmith 6 GB · worst run 1.1× +5%
- Avrea 8 GB · worst run 1.1× +10%
- Ubicloud standard · 8 GB · worst run 1.2× +11%
- GitHub 8 GB · worst run 1.3× +13%
- RunsOntmpfs r9g.large · 16 GB · worst run 1.6× +23%
- Warpbuild 8 GB · worst run 2.0× +39%
- Namespace 2x8 · 8 GB · worst run 1.2× +8% provisional
- NamespaceApple silicon 2x8 · 8 GB · worst run 1.2× +12% provisional
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
- WarpbuildApple silicon no 2 vCPU arm64 runner
How much a build's time moves from run to run on the same machine: the run that 9 runs in 10 beat (p90), over the typical run (p50), for each build on the machine Real builds picks, averaged over the builds with 3 runs or more. +10% means one run in ten took more than 10% longer than usual. A line measured on one build is shown, not ranked.
Predictabilityrun to runlower is betterarm64 · 4 vCPU Share Save image
RunsOn has the steadiest build times.
- RunsOn 2 machines · worst run 1.0× +2%
- Blacksmith 16 GB · worst run 1.1× +3%
- Warpbuild 16 GB · worst run 1.1× +4%
- RunsOntmpfs 2 machines · worst run 1.1× +4%
- GitHub 16 GB · worst run 1.1× +6%
- Avrea 16 GB · worst run 1.2× +10%
- Ubicloud standard · 16 GB · worst run 1.2× +12%
- Namespace 2 machines · worst run 1.3× +15%
- NamespaceApple silicon 4x16 · 16 GB · worst run 2.2× +31%
- WarpbuildApple silicon 6 vCPU · xfast (Apple M4 Pro) · 14 GB · worst run 2.2× +37% not ranked
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
How much a build's time moves from run to run on the same machine: the run that 9 runs in 10 beat (p90), over the typical run (p50), for each build on the machine Real builds picks, averaged over the builds with 3 runs or more. +10% means one run in ten took more than 10% longer than usual. A line measured on one build is shown, not ranked.
Predictabilityrun to runlower is betterarm64 · 8 vCPU Share Save image
RunsOn tmpfs and Blacksmith tie for the steadiest build times.
- RunsOntmpfs m9g.2xlarge · 32 GB · worst run 1.2× +5%
- Blacksmith 32 GB · worst run 1.1× +5%
- RunsOn 2 machines · worst run 1.1× +6%
- Warpbuild 32 GB · worst run 1.1× +6%
- Ubicloud standard · 32 GB · worst run 1.3× +12%
- Avrea 32 GB · worst run 1.3× +15%
- Namespace 2 machines · worst run 1.8× +16%
- WarpbuildApple silicon 12 vCPU · xfast (Apple M4 Pro) · 28 GB · worst run 2.6× +31% not ranked
- AWS CodeBuild no arm64 runner
- boat no arm64 runner
- GitHub no 8 vCPU arm64 runner
- NamespaceApple silicon no 8 vCPU arm64 runner
- Sengi no arm64 runner
- StarSling no arm64 runner
How much a build's time moves from run to run on the same machine: the run that 9 runs in 10 beat (p90), over the typical run (p50), for each build on the machine Real builds picks, averaged over the builds with 3 runs or more. +10% means one run in ten took more than 10% longer than usual. A line measured on one build is shown, not ranked.
Every benchmark
- Explorer Every benchmark, provider and runner in one place: pick a benchmark, open a provider, compare its runners.
- Builds in detail Every build phase, p50 to max, by best value, time or cost.
- Queue How long jobs wait for a runner in a burst, and how far build times move from run to run.
- Bigger runners What doubling the vCPU does to a build's time and cost.
- All runners All 102 runners in one table, and the price / performance chart.
- CPU Single and multi-thread scores, job-to-job spread, score per dollar.
- Cache Save and restore of a 4 GB cache, per cache action and storage.
- Disk I/O fio in the job's workspace: random and sequential, latency, sync writes.
- EC2 storage NVMe, gp3 and tmpfs on the same instances, and what each does to a build.
- EC2 CPU Every EC2 instance family by CPU score and price.
- Head-to-heads Two runners of the same size, build for build, cost first.
- Methodology How every figure is measured and ranked, and what changed.
- Calculator Your own minutes and sizes: what each provider would bill.