AWS AgentCore vs Google Agent Runtime
Both are hyperscaler platforms that host your agent code and give it places to run code and a browser. AgentCore runs each session in a microVM with root that stops after 8 hours, while Google runs sandboxes as an unprivileged user in a container that can live up to 7 days and be snapshotted with its memory.
The companies and products
AWS AgentCore
Amazon Bedrock AgentCore is a set of AWS services to deploy and run AI agents: a Runtime that hosts your agent code, plus built-in Code Interpreter and Browser tools.
AgentCore is not a sandbox product first. It is a hyperscaler's agent platform: you push a container image or code, and each Runtime session gets its own Firecracker microVM that you reach through AgentCore APIs. Around it sit managed tools (Code Interpreter, Browser) and services for memory, identity, gateways and evaluations, and an Instances compute type runs longer sessions on EC2 in your own account. It is built for teams who already build on AWS and want agents to stay inside their account, IAM and VPC.
Google Agent Runtime
Google Agent Runtime is the agent hosting part of Gemini Enterprise Agent Platform on Google Cloud, renamed from Vertex AI Agent Engine in April 2026.
It is a hyperscaler service, not a sandbox company. Agent Runtime deploys agents you build with ADK, LangGraph, LangChain, AG2 or LlamaIndex and bills them only while they work on a turn. Next to it, Agent Sandbox gives those agents a place to run code: a Code Execution sandbox, a shell or custom container sandbox, and a Computer Use browser. All of them run as secure containers inside your Google Cloud project.
Which use cases each one fits
For each use case, Battleships prices every sandbox provider on the same workload and sorts them from the cheapest monthly bill up. “7th cheapest of 345” means 6 of the other 344 providers run that workload, with every feature it needs, for less. “With limits” means the provider runs it but misses something, so it is listed after every full match. Open a use case for its whole ranking and what each provider misses.
| Use case and its workload | AWS AgentCore: price rank | Google Agent Runtime: price rank | bochka: price rank |
|---|---|---|---|
| Software factory 100 at once · 4 vCPU / 8 GB · 2,200 sessions of 8 h a month A coding agent needs root and Docker for most builds: AgentCore gives root but stops at 2 vCPU and 8 hours, Google gives neither root nor Docker, and bochka gives a full VM with both. | 100th of 367, with limits $888 a month | 295th of 367, with limits $5,801 a month | 3rd cheapest of 367 $634 a month |
| Always-on AI employees 50 machines 24/7 · 2 vCPU / 4 GB Google fits an assistant that lives for days better, since a sandbox can last up to 7 days, while AgentCore needs the Instances type on EC2 to pass 8 hours; bochka has no session cap. | 64th of 366, with limits $1,300 a month | 28th cheapest of 366 $7,519 a month | 5th cheapest of 366 $657 a month |
| Code interpreter bursts 100 at once · 1 vCPU / 2 GB · 200,000 sessions of 1 min a month This is AgentCore's home ground: its Code Interpreter bills only for CPU actually used, while Google's Code Execution sandbox is billed for as long as it exists. | 7th cheapest of 345 $114 a month | 254th of 345, with limits $275 a month | 186th of 345, with limits $2,000 a month |
| Sandboxes for your users 300 at once · 2 vCPU / 4 GB · 30,000 sessions of 23 min a month Neither gives your users a public URL for the app their sandbox runs; bochka serves HTTPS on up to 50 ports per sandbox. | 210th of 361, with limits $506 a month | 280th of 361, with limits $1,916 a month | Cheapest of 361 $209 a month |
| RL rollouts 1,000 at once · 2 vCPU / 4 GB · 1,000,000 sessions of 3 min a month | 125th of 360, with limits $8,075 a month | 128th of 360, with limits $8,240 a month | 34th of 360, with limits $2,000 a month |
| Agent evals 250 at once · 4 vCPU / 8 GB · 15,000 sessions of 20 min a month | 33rd cheapest of 364 $650 a month | 79th cheapest of 364 $2,060 a month | 10th cheapest of 364 $200 a month |
| CI runners 40 at once · 8 vCPU / 16 GB · 6,000 sessions of 12 min a month CI jobs that build containers need Docker, which Google's unprivileged sandboxes do not offer and AgentCore does not document; bochka runs Docker and /dev/kvm on every VM. | 52nd cheapest of 377 $399 a month | 98th cheapest of 377 $989 a month | 4th cheapest of 377 $86 a month |
| 24/7 dev boxes 1 machine 24/7 · 4 vCPU / 8 GB Neither has SSH or a disk that keeps everything for months, so neither works as a box you log into; bochka gives you SSH, a desktop and a disk saved every minute. | 228th of 368, with limits $52 a month | 311th of 368, with limits $241 a month | 18th cheapest of 368 $26 a month |
| Computer-use agents 50 at once · 4 vCPU / 8 GB · 20,000 sessions of 15 min a month Both give you a browser rather than a whole desktop; bochka streams a full Linux desktop at 60fps for agents that use other apps too. | 104th of 369, with limits $202 a month | 283rd of 369, with limits $1,648 a month | 6th cheapest of 369 $200 a month |
| Browser-use agents 50 at once · 2 vCPU / 4 GB · 30,000 sessions of 10 min a month AgentCore Browser is a strong fit here, with live view, session replay, profiles and proxies; Google's Computer Use browser offers a VNC view and CDP, with outbound internet off unless turned on. | 4th cheapest of 367 $144 a month | 102nd of 367, with limits $824 a month | 7th cheapest of 367 $200 a month |
| Scraping & bots (startup scale) 50 at once · 1 vCPU / 2 GB · 100,000 sessions of 3 min a month | 27th cheapest of 346 $287 a month | 53rd cheapest of 346 $515 a month | 6th cheapest of 346 $90 a month |
| Mass scraping & bots (undetectable) 1,000 at once · 1 vCPU / 2 GB · 3,000,000 sessions of 2 min a month | 165th of 338, with limits $5,458 a month | 210th of 338, with limits $8,240 a month | 71st of 338, with limits $1,800 a month |
| Parallel red teaming 200 at once · 2 vCPU / 4 GB · 5,000 sessions of 30 min a month | 62nd of 360, with limits $294 a month | 260th of 360, with limits $412 a month | 2nd cheapest of 360 $200 a month |
| GPU eval jobs 10 at once · 8 vCPU / 32 GB · 500 sessions of 30 min a month | Can't run it | Can't run it | Can't run it |
| macOS agents 5 at once · 4 vCPU / 8 GB · 1,000 sessions of 20 min a month | Can't run it | Can't run it | Can't run it |
Key features
| Feature | AWS AgentCore | Google Agent Runtime | bochka |
|---|---|---|---|
| Isolation | Firecracker microVM per Runtime session; Code Interpreter and Browser in a containerized environment | Secure container sandboxing inside your Google Cloud project | Full VM with its own kernel on dedicated AMD Ryzen 9 9950X hosts |
| Root, Docker and /dev/kvm | Root inside the microVM; arm64 container images; Docker and /dev/kvm not documented | Runs as appuser with no sudo; images must not need root; no Docker or /dev/kvm | Root, systemd, Docker and /dev/kvm on every machine |
| How long a session lives | Up to 8 hours on microVMs, stopped after 15 idle minutes by default; up to 14 days on Instances | TTL of up to 7 days for a custom container and 14 days for Code Execution, then deleted | No session cap on paid plans; you set the time to live |
| Files after a stop | Only one storage mount under /mnt, up to 1 GB, in preview; EBS volumes on Instances; the rest resets | Pause keeps the file system; everything is deleted when the TTL ends | The whole disk, saved every minute and on stop; a stopped sandbox costs nothing |
| Snapshots and branching | No user snapshots or fork of a session | Snapshots keep memory and files; start many sandboxes from one snapshot | Automatic disk snapshots every minute; fork a running sandbox with its disk (no memory snapshots) |
| Machine size | 2 vCPU / 8 GB per microVM session, not adjustable; bigger EC2 types only on Instances | Runtime: 1 to 8 vCPU and up to 32 GiB; sandbox maximum not published | Up to 8 vCPU / 16 GB on the $20 plan, 16 vCPU / 32 GB from the $100 plan |
| Reaching the sandbox | No open ports or SSH; only through AgentCore APIs with IAM or JWT | Through Google's API or a private endpoint in your VPC; no public URL or SSH | HTTPS URLs for up to 50 ports, token-protected by default; raw TCP; SSH |
| Network controls | VPC mode with security groups, PrivateLink and a NAT gateway for a fixed outbound IP | Outbound internet off unless the template turns it on; Cloud NAT, static IP, Cloud NGFW or Secure Web Proxy | Open internet on every plan, no provider-side allowlist |
| Browser and desktop | Managed Chromium with live view, session recording and replay, profiles and proxies | Browser in a container with a VNC live view, a click, type and screenshot API, and CDP | Full Linux desktop streamed at 60fps, with Chromium, computer use and browser control |
| Coding agents | AWS's managed harness built on Strands Agents, or your own agent code | Managed Agents API (Preview) runs Google's Antigravity harness | Claude Code, Codex, OpenCode and more preinstalled, with a prompt, steer and events API |
| Hosting your agent code | Runtime hosts your container or code in any framework, one microVM per session | Agent Runtime deploys ADK, LangGraph, LangChain, AG2 and LlamaIndex agents; idle time between turns is not billed | You run your own agent loop inside a sandbox |
| Keeping API keys out of the sandbox | AgentCore Identity stores OAuth tokens and API keys and hands them to your agent code | You provision and scope the credentials the agent uses | Credentials proxy for model providers and any HTTPS API; switch a credential to another account when credits run out |
| GPUs | NVIDIA G-series and Inferentia2 on the Instances type | No | No |
| Regions | 17 commercial AWS regions plus GovCloud | 23 locations in the US, Europe, Asia, Canada, Australia and Brazil | EU only: Germany, Finland, France |
Every feature Battleships tracks (43)
| Feature | AWS AgentCore | Google Agent Runtime | bochka |
|---|---|---|---|
| Full VM (own kernel) | Yes | No | Yes |
| Docker inside | Not published | No | Yes |
| Nested virtualization | No | No | Yes |
| Root | Yes | No | Yes |
| Systemd | No | Not published | Yes |
| Desktop GUI | No | No | Yes |
| Browser | Yes | Yes | Yes |
| Computer-use API | Yes | Yes | Yes |
| SSH | No | No | Yes |
| Persistent disk | Yes | Yes | Yes |
| Pause/resume | No | Yes | Yes |
| Automatic snapshots | No | Yes | Yes |
| Fork/clone | No | Yes | Yes |
| Live fork (no pause) | No | Yes | Yes |
| Memory fork | No | Yes | No |
| Custom image (Docker or snapshot) | Yes | Yes | Yes |
| ≥24 h sessions | No | Yes | Yes |
| Idle auto-stop | Yes | No | No |
| Wake on request | Yes | No | No |
| HTTPS preview URLs | No | No | Yes |
| Raw TCP inbound | No | Not published | Yes |
| Public IPv4 | No | No | No |
| Egress allowlist | Yes | Yes | No |
| Secret proxy | No | No | Yes |
| Secret proxy for any API | No | No | Yes |
| Preinstalled agents | No | Yes | Yes |
| Agent harness API | No | Yes | Yes |
| MCP server | Yes | No | No |
| GPU | Yes | No | No |
| Windows | No | No | No |
| Arm64 | Yes | Yes | No |
| Self-hosting | No | No | No |
| BYOC | Yes | No | No |
| Open source | No | No | No |
| Scoped API keys | Not published | Not published | Yes |
| Spend limits | Yes | Not published | Yes |
| Usage per sandbox (API) | No | Yes | Yes |
| Audit logs | Yes | Not published | No |
| SSO | Yes | Yes | No |
| SOC 2 | Yes | Yes | No |
| HIPAA | Yes | Yes | No |
| EU data residency | Yes | Yes | Yes |
| Zero data retention | Not published | Not published | Yes |
Performance
There is no comparable measurement for AWS AgentCore or Google Agent Runtime: neither was part of the hpc-sandbox-benchmarks run. bochka did 26.69 Node.js build-and-test runs per second at 4 vCPU / 8 GB in StarSling's run of 2026-09-29. An AgentCore microVM stops at 2 vCPU, so a like-for-like test there would need the Instances type.
Node.js build-and-test loop at 4 vCPU / 8 GB, higher is better. Measured with the open hpc-sandbox-benchmarks suite. AWS AgentCore and Google Agent Runtime have no published run in this suite yet, so there is no like-for-like number to compare.
Pricing in five scenarios
Each bill is priced from the providers' public price lists by the open Battleships engine, with the cheapest plan that runs the scenario.
| Scenario | AWS AgentCore, a month | Google Agent Runtime, a month | bochka, a month |
|---|---|---|---|
| One coding agent, workdays 4 vCPU / 8 GB, 50 GB disk, 8 hours a day, 22 days a month, files kept between days | $26 · Pay-as-you-go (Instances) · with limits | $73 · Pay as you go | $20 · $20/mo |
| One agent that never stops 4 vCPU / 8 GB, 50 GB disk, running 24/7 all month | $107 · Pay-as-you-go (Instances) · with limits | $301 · Pay as you go | $26 · $20/mo |
| 50 agents in parallel 50 machines of 4 vCPU / 8 GB at once, 8-hour sessions, 1,100 sessions a month, Docker inside | $1,286 · Pay-as-you-go (Instances) · with limits | $3,626 · Pay as you go · with limits | $317 · $20/mo × 3 seats |
| Sandboxes for your users 2 vCPU / 4 GB, 30,000 sessions of 10 minutes a month, up to 300 at once | $423 · Pay-as-you-go · with limits | $1,030 · Pay as you go | $200 · $100/mo × 2 seats |
| 1,000 short jobs at once 2 vCPU / 4 GB, 1,000 at once, 3 minutes each, 200,000 jobs a month | $1,615 · Pay-as-you-go (Instances) | $2,060 · Pay as you go | $2,000 · $2000/mo · with limits |
AWS AgentCore: Runtime microVMs, Code Interpreter and Browser bill per second for CPU actually used and for peak memory. The 4 vCPU shape in these scenarios is the Instances type: the EC2 on-demand price plus a 12% management fee, with storage and data transfer billed on top.
Google Agent Runtime: Google bills $0.085 per vCPU-hour and $0.009 per GiB-hour allocated, with the first 50 vCPU-hours and 100 GiB-hours free each month. Runtime skips idle time between turns, but sandboxes are billed while they exist, and disk and egress rates are not published.
Google Agent Runtime: Plans start at $20 a month, and the whole fee is usage credit spent per second on running sandboxes. A stopped sandbox costs nothing and keeps its disk.
Limits in those scenarios
AgentCore is lower in the table in all five scenarios, but it fits only the 1,000-job burst: the others need 4 vCPU or more than 8 hours, which pushes them onto Instances, where files outside the storage mount still reset, and its Code Interpreter is a different kind of product from sandboxes for your users. Google fits four of the five, but its 50-agent fleet fails because there is no Docker inside, and every Google row leaves disk and egress unpriced. bochka fits four of five; at 1,000 short jobs at once its $2000 plan queues starts above 90 per minute.
| Scenario | AWS AgentCore | Google Agent Runtime | bochka |
|---|---|---|---|
| One coding agent, workdays |
|
| None that bite |
| One agent that never stops |
|
| None that bite |
| 50 agents in parallel |
|
| None that bite |
| Sandboxes for your users |
|
| None that bite |
| 1,000 short jobs at once |
|
|
|
Which one should you use?
Use AWS AgentCore when
your agents must stay inside your AWS account and VPC, need root or a GPU, or are short code and browser tasks that fit the managed tools.
- Root inside each session. Agents that need to install packages as root get it in their own microVM, which Google's unprivileged sandboxes do not allow.
- GPUs and two-week sessions on Instances. Teams on AWS can run agents on EC2 in their own account for up to 14 days, with NVIDIA or Inferentia2 GPUs.
- Managed tools billed only for what they use. For short code runs and browser tasks, Code Interpreter and Browser charge for CPU actually used, with live view and replay.
Use Google Agent Runtime when
you build agents with ADK or LangGraph on Google Cloud and want sandboxes that last for days, memory snapshots and egress locked down by default.
- Sandboxes that last for days. A custom container sandbox can live up to 7 days without extra setup, where an AgentCore microVM stops after 8 hours.
- Memory snapshots and branching. Teams that want to try many paths from one state can snapshot a sandbox with its memory and start many copies.
- A managed home for ADK and LangGraph agents. If you build with ADK or LangGraph on Google Cloud, Agent Runtime hosts them and skips billing for idle time between turns.
Where bochka fits
AgentCore and Agent Runtime are best when your agents must live inside AWS or Google Cloud and mostly run short code or browser tasks. If your agents do real work on a machine for hours or days, installing tools, running Docker, building apps and showing them to people, bochka gives each one a whole VM that keeps its disk and has no session cap on paid plans.
A whole machine with Docker and root
Google sandboxes run without root or Docker, and AgentCore does not document Docker and stops at 2 vCPU per microVM. Bigger AgentCore machines mean running EC2 in your own account. bochka has root, Docker, systemd and /dev/kvm on every VM.
A disk that stays, with fork
On AgentCore only a 1 GB mount survives a stop, so setup repeats on every resume. Google keeps files until the TTL ends, then deletes the sandbox. bochka saves the whole disk every minute and can fork a running sandbox.
Open the app your agent is building
Neither gives a sandbox a public URL: AgentCore is reached only through its APIs, and Google through its API or a private VPC endpoint, so you build your own proxy. bochka serves HTTPS on up to 50 ports, plus raw TCP and SSH.
Use bochka when your agents do real work on a machine for hours: installs, builds, tests, Docker, emulators and a desktop, with a disk that keeps everything and a URL you can open. Compare it one to one: bochka vs AWS AgentCore and bochka vs Google Agent Runtime.
Questions people ask
Is AWS AgentCore cheaper than Google Agent Runtime?
In our five scenarios AgentCore's numbers are lower every time, for example $1,615 against $2,060 for 1,000 short jobs at once. But AgentCore fits only that one scenario, while Google fits four. bochka comes to $20 a month for one coding agent on workdays, against $26 on AgentCore and $73 on Google.
Which is faster, AgentCore or Google Agent Runtime?
Neither has been measured in the hpc-sandbox-benchmarks run, so there is no fair answer yet. bochka did 26.69 runs per second on the Node.js build-and-test loop at 4 vCPU / 8 GB.
Can AgentCore or Google Agent Runtime run Docker?
AgentCore runs your arm64 container image as root in a microVM, but its docs do not say Docker works inside a session. Google sandboxes run as the unprivileged user appuser with no sudo, so Docker is not available. bochka runs Docker and exposes /dev/kvm on every VM.
How long can a session run on each?
An AgentCore microVM session lasts up to 8 hours and stops after 15 idle minutes by default; the Instances type runs up to 14 days. A Google sandbox lives until its TTL ends, up to 7 days for a custom container and 14 days for Code Execution. bochka has no session cap on paid plans.
Is Google Agent Runtime the same as Vertex AI Agent Engine?
Yes. Google renamed Vertex AI Agent Engine to Agent Runtime in April 2026, as part of Gemini Enterprise Agent Platform. Its code, shell and browser sandboxes are a separate part of the same platform called Agent Sandbox.
Is bochka SOC 2 compliant?
SOC 2 is in progress. We're open about where we stand: live status at status.bochka.xyz and our data retention policy at docs.bochka.xyz/data-retention.
Try bochka
Install the CLI, then run bochka new.
Product preview
More comparisons
- bochka vs AWS AgentCore
- bochka vs Google Agent Runtime
- E2B vs AWS AgentCore
- Daytona vs AWS AgentCore
- Modal vs AWS AgentCore
- Vercel Sandbox vs AWS AgentCore
- Cloudflare Sandbox vs AWS AgentCore
- Blaxel vs AWS AgentCore
- Runloop vs AWS AgentCore
- Tensorlake vs AWS AgentCore
- Morph Cloud vs AWS AgentCore
- CodeSandbox SDK vs AWS AgentCore