RunsOn m8i-flex.xlarge · EBS gp3 vs Blacksmith 4 vCPU x64
RunsOn does each build for 2.1–3.8× less; on time, Blacksmith is faster on the Docker build and neither is clearly faster on the Rust and TypeScript builds. Both 4 vCPU x64.
All three builds 2.4× cheaper on RunsOn: $0.0309 vs $0.0741Blacksmith 1.9× faster: 8m 27s vs 16m 06s
Rust build 2.9× cheaper on RunsOn: $0.0054 vs $0.0155within run-to-run spread: 2m 49s · 1m 47s
TypeScript build 3.8× cheaper on RunsOn: $0.0032 vs $0.0121within run-to-run spread: 1m 31s · 1m 19s
Docker build 2.1× cheaper on RunsOn: $0.0224 vs $0.0465Blacksmith 2.2× faster: 5m 22s vs 11m 46s
RunsOn m8i-flex.xlarge · EBS gp3
jobs ran from Amazon.com (US)
Blacksmith 4 vCPU
jobs ran from Secured Servers (US)
Side by side
Every measurement both runners share, cost first. The dot marks the better value; gap is the larger over the smaller.
| Measurement | RunsOn | Blacksmith | Gap |
|---|---|---|---|
| Cost | |||
| Cost per Rust build | $0.0054 | $0.0155 | 2.9× RunsOn |
| Cost per TypeScript build | $0.0032 | $0.0121 | 3.8× RunsOn |
| Cost per Docker build | $0.0224 | $0.0465 | 2.1× RunsOn |
| Price per minute | $0.0012 | $0.0080 | 6.4× RunsOn |
| Real builds, wall time | |||
| Rust build (spotify-player) | 2m 49s | 1m 47s | ≈ even |
| TypeScript build (opencode) | 1m 31s | 1m 19s | ≈ even |
| Docker build (PostHog) | 11m 46s | 5m 22s | 2.2× Blacksmith |
| CPU & memory | |||
| PassMark single-thread (points) | 3,267 | 4,332 | 1.3× Blacksmith |
| PassMark CPU Mark, all cores (points) | 8,030 | 14,771 | 1.8× Blacksmith |
| sysbench CPU, 1 thread (events/s) | 3,054 | 6,493 | 2.1× Blacksmith |
| 7-Zip compression (MIPS) | 23,275 | 28,802 | 1.2× Blacksmith |
| Memory read bandwidth | 21.8 GiB/s | 41.7 GiB/s | 1.9× Blacksmith |
| Measurement | RunsOn | Blacksmith | Gap |
|---|---|---|---|
| Workspace disk (fio) | |||
| Random read, 4k QD32 (IOPS) | 2,993 | 728,035 | 243.3× Blacksmith |
| Random write, 4k QD32 (IOPS) | 2,994 | 218,915 | 73.1× Blacksmith |
| Random read, 4k QD1 (IOPS) | 1,753 | 71,728 | 40.9× Blacksmith |
| QD1 read latency, p99 | 918 µs | 16 µs | 57.7× Blacksmith |
| Sequential read | 401 MiB/s | 7.0 GiB/s | 18.0× Blacksmith |
| Sequential write | 401 MiB/s | 1.7 GiB/s | 4.3× Blacksmith |
| Sync write, 4k + fdatasync (IOPS) | 525 | 17,925 | 34.2× Blacksmith |
| Network, cache & Docker | |||
| Download from GitHub | 135 MiB/s | 76 MiB/s | 1.8× RunsOn |
| Cache restore, 4 GB (actions/cache) | 17s | 11s | 1.5× Blacksmith |
| Cache save, 4 GB (actions/cache) | 14s | 7s | 2.2× Blacksmith |
| docker pull | 6.0s | 7.9s | 1.3× RunsOn |
| Database | |||
| pgbench read/write (TPS) | 3,062 | 12,408 | 4.1× Blacksmith |
| pgbench read-only (TPS) | 55,039 | 109,979 | 2.0× Blacksmith |
| Queue | |||
| Queue under burst load | 17s | 10s | 1.7× Blacksmith |
- ≈ even: under 5% apart, or a build inside both runners' run-to-run spread
- Queue: identical jobs queued at the same instant (burst suite)
- RunsOn at EC2 spot, its default (7-day average of the cheapest us-east-1 zone's spot price, fetched 16 Nov 2026); 3 of its jobs fell back to on-demand (AWS had no spot capacity for the type) and are priced at the spot rate over the same billed time
- Cost: a managed SaaS runner against RunsOn self-hosted on your AWS account; the effort of running it was not measured