Skip to content
Provider directory
Tensorlake logoTensorlake

Stateful microVMs that suspend on idle and wake in a second.

Run in a trial

Objective

95.6

speed · reliability · cost

Measured runs

3

real samples

Median

895 ms

total wall time

p95

tail latency

Success

100.0%

completed runs

Avg cost

$0.00000

per run

Runtime

microvm

Firecracker / Cloud Hypervisor

State & compute

Stateful · Full Linux

CPU only

Commercial shape

$0 idle

per-second, suspend stops meter

Per workload

Where the latency comes from

Center and tail stay together, with the sample and failure rate visible.

WorkloadnMedianp99ColdCostSuccess
Cold start3895 ms939 ms788 ms$0.00000100.0%

Recent evidence

Individual measured runs

The latest runs are useful for spotting changes that a long-run aggregate can hide.

WorkloadOpponentTotalProvisionCostStatus
Cold startDaytona891 ms775 ms$0.00000Pass
Cold startDaytona940 ms809 ms$0.00001Pass
Cold startDaytona895 ms788 ms$0.00000Pass