AI sandbox benchmark
Last updated 2026-07-26 · figures are vendor-published; corrections welcome
The cold-start race
Bar length is the published start-up time, to scale. The gap is the story: Blaxel at 25ms is roughly 32x quicker off the line than the slowest here.
- 25ms
- 50ms
- 90ms
- 150ms
- 300ms
- 300ms
- 400ms
- 400ms
- 500ms
- 800ms
Not strictly like-for-like: some figures are resume-from-hibernation or V8-isolate starts, not a full cold boot. The full table and method are below.
| Sandbox | Isolation | Cold start | $/vCPU-hr | GPU | Session cap |
|---|---|---|---|---|---|
| Blaxel | microVM | 25ms | $0.083 | No | Persistent (hibernation) |
| Cloudflare Sandbox SDK | V8 isolate + container | 50ms | $0.072 | No | 30min execution cap |
| Daytona | Docker container | 90ms | $0.05 | Yes | Persistent (lifecycle-managed) |
| E2B | Firecracker microVM | 150ms | $0.05 | No | Up to 24h |
| Northflank | Kata / gVisor microVM (selectable) | 300ms | $0.01667 | Yes | Unlimited |
| Fly.io Machines | microVM | 300ms | $0.07 | Yes | Persistent |
| Vercel Sandbox | Firecracker microVM | 400ms | $0.128 | No | 45min (Hobby) / 24h (Pro) |
| Runloop | microVM | 400ms | $0.108 | No | Configurable |
| CodeSandbox SDK | Firecracker microVM | 500ms | $0.05 | No | Configurable |
| Modal | gVisor | 800ms | $0.14 | Yes | Configurable |
Method: values are the providers' own published start-up and pricing figures as of 2026-07-26, normalized to milliseconds and USD/vCPU-hour. The start-up figures are not strictly like-for-like: Blaxel's ~25ms is a resume from hibernation and Cloudflare's ~50ms is a V8-isolate start, versus a full microVM/container cold boot elsewhere. Start-up time also varies with image size, region and warm-pool state, so treat these as directional. Independent measured runs are on the roadmap; this table will carry them when live.
How to read this
Cold start and headline price are the two numbers buyers anchor on, but they rarely decide the bill alone. Isolation strength (microVM vs container), idle/paused cost, session caps, GPU availability and creation fees all move real spend. Use this table to shortlist, then model your actual workload in the cost calculator.
FAQ
Which AI sandbox has the fastest cold start?
As of 2026-07-26, Blaxel has the fastest published start-up figure in our benchmark at 25ms, using microVM. Note this is a resume-from-hibernation time rather than a cold boot for some providers, so compare the measurement basis on each provider page before ranking on the raw number.
Which AI sandbox is cheapest per vCPU-hour?
Northflank has the lowest published rate at $0.01667 per vCPU-hour, though total cost depends on idle time and session length.
What moved in the sandbox world, every Friday.
Pricing changes, new entrants, benchmark refreshes and one sharp take from Eve Harper. Read by engineers choosing where their agents and apps run. No fluff. Read the archive.