CapabilitiesLifecycle
Starts in under a second
The provider documents new or resumed machines being ready in under a second.
Short answer: 10 of 16 providers offer it fully: Blaxel, Daytona, Lizard, Modal, Runtime, Sandboxer, Sprites, Tensorlake, Vercel Sandbox and boxd. With a catch: InstaBox. Checked 29 October 2026.
Who offers it
Part of Lifecycle: stop, pause and sleep, which explains why it matters.
Fully: Blaxel, boxd, Daytona, Lizard, Modal, Runtime, Sandboxer, Sprites, Tensorlake and Vercel Sandbox. With a catch: InstaBox. Not at all: Boat.
That is the resume from standby; a brand-new sandbox still has to start.
docs-confirmed “Sandboxes resume from standby in under 25 milliseconds and automatically scale to zero after a few seconds of inactivity.” Source
That is the wake from standby; waking from hibernation takes about 85 ms.
docs-confirmed “The first inbound packet (an SSH connection, an HTTP request, even a single TCP SYN) wakes it in **sub-millisecond** time.” Source
The claim is for creating a sandbox from code to execution.
docs-confirmed “Sub 90ms sandbox creation from code to execution” Source
Measured from create to first Python output on 2026-10-12.
docs-confirmed “the median time from creating a sandbox to its first Python output was 228 ms” Source
The claim is on the provider’s front page, not in the docs.
docs-confirmed “Sub-second starts on large custom images and at high concurrency.” Source
The figure is for waking a paused machine; a brand-new sandbox still has to boot.
docs-confirmed “await spare.wake(); // about 0.1 s” Source
Measured on 2026-10-28 from US-East for the default size.
docs-confirmed “One create at a time (10 each, median / slowest) | 0.18 s / 0.32 s” Source
A cold wake takes 1–2 s instead.
docs-confirmed “The next request resumes it in 100–500ms, and processes pick up mid-thought, exactly where they were.” Source
A pool claim is faster still than a cold start.
docs-confirmed “Tensorlake provides isolated MicroVM sandboxes that boot in hundreds of milliseconds” Source
The claim is a line in the concepts table.
docs-confirmed “Startup time| Sub-second| Milliseconds (Firecracker optimized for fast boot)” Source
The catch: A resume usually takes about a second, which is not under one.
docs-confirmed “// carries on where it left off, usually in about a second” Source
Note: A resume or a new machine takes a few seconds, not under one.
docs-confirmed “A resume, a fork or a deploy is ready in **a few seconds, whatever the size of the disk**.” Source
- Agent 37Not published
No start or resume time is published anywhere in the docs.
not-published “Wake it with [`start`](/docs/agents-api/instances/start), a command sent through [exec](/docs/agents-api/exec), or any request to its URLs.” Source
- Cloudflare SandboxNot published
No start or resume time is published.
not-published “An instance runs only until something stops it, and its processes and the files on its disk end with it. The next `start()` creates a new instance.” Source
- E2BNot published
No start or resume time is published in the docs.
not-published “Sandboxes stay running as long as you need them.” Source
- MorphNot published
No start or resume time is published.
not-published “Each instance runs from a saved snapshot.” Source
Related
Need several capabilities at once? Add more in the capability finder. Something wrong here? Use the report link at the bottom of the page.