AWS AgentCore Runtime

https://aws.amazon.com/bedrock/agentcore/

AWS Bedrock AgentCore Runtime hosts AI agents and tools in managed compute sessions. It offers serverless microVM sessions and managed EC2-backed Instances; Instances support persistent multi-day sessions, GPU workloads, and multiple agents on a shared instance.

Pricing model Consumption-based; per-second billing on actual CPU and memory usage, with a 1-second minimum and 128 MB minimum memory billing.
Pricing The headline compute price as the provider states it. Sorting ranks prices stated per vCPU over time, converted to hourly. Other pricing models sort after, unranked. $0.1276 per vCPU-hour for Runtime V2 consumption pricing
Free while idle Whether a stopped, paused, or sleeping sandbox costs nothing. ✕
Elastic Resources flex with what the sandbox actually uses, and the bill follows: yes when the provider bills on active or observed use, no when capacity is reserved or allocated and billed while it runs. ✓
Memory snapshots Whether RAM state survives a pause and resume, not just disk. ✕
Wake on request Whether a stopped sandbox wakes automatically on inbound traffic. ✓
Start / resume Typical time from create or resume to running. Sorting ranks stated times, converted to seconds; claims with no figure sort after, unranked. P75 1.9–2.0 seconds for 200 MB–2 GB container images on Runtime V2
Max runtime The longest a sandbox may run. Up to 8 hours per microVM lifecycle; up to 14 days on Instances.
Isolation The isolation technology between sandboxes. Dedicated microVM per user session for serverless compute, with hardware-enforced session isolation; Instances use managed EC2 compute.
GPUs Whether GPU instances are available for sandboxes. ✓
Docker Whether Docker containers can run inside the sandbox. ✓
SDKs Python, TypeScript

✓ yes · ✕ no · – no cited public fact.

Capabilities depend on the compute type: microVM sessions run for up to 8 hours per lifecycle, while Instances can run for up to 14 days and provide GPU support. Persistent filesystem state is documented, but RAM snapshots across stop/resume are not. V1 remains the default platform version, and AWS bills it at $0.0895 per vCPU-hour, less than the V2 consumption rate. For microVMs, AWS charges for the span from ready, through idle periods, to termination, plus system overhead and a 128 MB memory floor.

Sources

Last verified 2026-10-03. Corrections land through the update log.