bochka vs Cerebrium

These are different kinds of product. bochka gives each agent its own full Linux machine to work inside, while Cerebrium is a serverless GPU cloud that runs your model or app as containers that scale up with requests and down to zero.

Where bochka stands out

A whole machine to work inside

Each bochka sandbox is a full VM with its own kernel, root, SSH, systemd and Docker, on dedicated AMD Ryzen 9 9950X hosts. It also exposes /dev/kvm, so Android emulators and VMs run inside it. Cerebrium runs one app container that you reach through its endpoints, with no shell to log into.

The whole disk stays, and you can fork it

bochka saves the disk every minute and when the sandbox stops, and a stopped sandbox costs nothing and keeps every file. You can also fork a running sandbox into a copy with the same disk. On Cerebrium only the /persistent-storage folder survives a restart, and there is no fork.

Agents, desktop and keys built in

Claude Code, Codex and five other agents come preinstalled, and you prompt, steer and follow them through one API. Each sandbox has a 60fps desktop with Chromium for computer use. Real API keys stay outside the sandbox in a credentials proxy that works with any HTTPS API.

The companies and products

bochka

bochka.xyz runs sandboxes for AI agents: each one is a full Linux VM with its own kernel.

You start a machine with one command or one API call, then SSH in, run Docker, open a desktop, or let a built-in agent (Claude Code, Codex, OpenCode and others) work in it. The disk stays between sessions, and a stopped sandbox costs nothing.

Isolation: Full VM per sandbox, own kernel
Hardware: Dedicated AMD Ryzen 9 9950X (Zen 5) hosts
Regions: EU: Germany, Finland, France
Plans: From $20 a month, all of it usage credit, 7-day trial

Cerebrium

Cerebrium is a serverless GPU cloud for real-time AI: voice agents, video models, LLMs and other ML apps.

You describe an app in a cerebrium.toml file, with your Python code or your own Dockerfile, and Cerebrium builds it and serves it as HTTP or WebSocket endpoints. It starts containers when requests arrive, scales them across GPUs and regions, and scales them back to zero when traffic stops. It is built for teams that serve models in production, not for agents that need a machine of their own, so this page compares the places where the two overlap.

Company: Founded 2021, Y Combinator Winter 2022, based in New York City
Product: Serverless GPU inference: apps run as autoscaling containers behind HTTP or WebSocket endpoints
Isolation: Containers built from your code or Dockerfile; the docs do not name the isolation layer
Hardware: CPU, T4, A10, L4, L40S, A100, H100, H200, B200 and RTX PRO 6000, up to 8 GPUs per container, plus AWS Inferentia2 and Trainium
Regions: US East, US Central, Stockholm and Finland generally available; 8 more, including Tokyo, Mumbai and Sao Paulo, on request
Plans: Hobby is free plus compute (500 CPU containers, 5 GPUs at once); Standard is $100 a month plus compute (1,000 containers, 30 GPUs); Enterprise is custom
Compliance: SOC 2, HIPAA, GDPR and ISO 27001 listed on the pricing page

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. “3rd cheapest of 367” means 2 of the other 366 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	bochka: price rank	Cerebrium: price rank
Software factory 100 at once · 4 vCPU / 8 GB · 2,200 sessions of 8 h a month A software factory needs Docker, root and a whole disk that is still there tomorrow; Cerebrium keeps only its /persistent-storage folder and has no documented Docker inside.	3rd cheapest of 367 $634 a month	182nd of 367, with limits $2,008 a month
Always-on AI employees 50 machines 24/7 · 2 vCPU / 4 GB An assistant that lives on one machine all month keeps its whole disk on bochka; on Cerebrium only one shared folder survives a restart, and the container keeps billing while it runs.	5th cheapest of 366 $657 a month	221st of 366, with limits $2,534 a month
Code interpreter bursts 100 at once · 1 vCPU / 2 GB · 200,000 sessions of 1 min a month Short code runs fit Cerebrium's model: a container starts on a request, bills per second, and scales back to zero when the work stops.	186th of 345, with limits $2,000 a month	225th of 345, with limits $131 a month
Sandboxes for your users 300 at once · 2 vCPU / 4 GB · 30,000 sessions of 23 min a month Both can serve your users; bochka gives each user a full VM with a lasting disk and hosting, while Cerebrium gives them a stateless endpoint that scales.	Cheapest of 361 $209 a month	230th of 361, with limits $821 a month
RL rollouts 1,000 at once · 2 vCPU / 4 GB · 1,000,000 sessions of 3 min a month	34th of 360, with limits $2,000 a month	72nd of 360, with limits $4,068 a month
Agent evals 250 at once · 4 vCPU / 8 GB · 15,000 sessions of 20 min a month	10th cheapest of 364 $200 a month	32nd cheapest of 364 $639 a month
CI runners 40 at once · 8 vCPU / 16 GB · 6,000 sessions of 12 min a month CI jobs that build containers or Android apps need Docker and /dev/kvm, which every bochka sandbox has and Cerebrium does not document.	4th cheapest of 377 $86 a month	81st cheapest of 377 $766 a month
24/7 dev boxes 1 machine 24/7 · 4 vCPU / 8 GB A dev box needs SSH and a disk that keeps your installs; Cerebrium has no SSH and resets everything outside /persistent-storage.	18th cheapest of 368 $26 a month	259th of 368, with limits $102 a month
Computer-use agents 50 at once · 4 vCPU / 8 GB · 20,000 sessions of 15 min a month bochka ships a Linux desktop streamed at 60fps with Chromium for computer use; Cerebrium has no desktop or browser product.	6th cheapest of 369 $200 a month	179th of 369, with limits $575 a month
Browser-use agents 50 at once · 2 vCPU / 4 GB · 30,000 sessions of 10 min a month	7th cheapest of 367 $200 a month	221st of 367, with limits $287 a month
Scraping & bots (startup scale) 50 at once · 1 vCPU / 2 GB · 100,000 sessions of 3 min a month	6th cheapest of 346 $90 a month	223rd of 346, with limits $195 a month
Mass scraping & bots (undetectable) 1,000 at once · 1 vCPU / 2 GB · 3,000,000 sessions of 2 min a month	71st of 338, with limits $1,800 a month	135th of 338, with limits $3,991 a month
Parallel red teaming 200 at once · 2 vCPU / 4 GB · 5,000 sessions of 30 min a month	2nd cheapest of 360 $200 a month	197th of 360, with limits $191 a month
GPU eval jobs 10 at once · 8 vCPU / 32 GB · 500 sessions of 30 min a month Cerebrium has GPUs from T4 to B200 and up to 8 in a container; bochka is CPU only.	Can't run it	28th cheapest of 98 $1,975 a month
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

Key features

Feature	bochka	Cerebrium
What you get	A full VM with its own kernel on dedicated AMD Ryzen 9 9950X hosts, which you or an agent work inside	An app you deploy as containers built from your code or Dockerfile, served as endpoints
Shell access	SSH, root and systemd, like a normal Ubuntu server	No SSH; you reach the app through its HTTP or WebSocket endpoints
Docker and nested virtualization	Docker inside, and /dev/kvm for Android emulators and VMs	The app is the container; Docker inside and /dev/kvm are not documented
Files after a stop	The whole disk is saved every minute and on stop, and a stopped sandbox keeps it for free	Only the shared /persistent-storage volume survives; packages and files elsewhere reset
Fork and snapshots	Fork a running sandbox with its disk	No fork; an early-beta memory checkpoint (CPU and GPU memory, no files) speeds up new containers of the same build
Run time	No session cap on paid plans; you set the time to live	No container cap, but a request times out at response_grace_period (15 minutes by default, 12 hours at most for async runs)
Scale to zero and wake	No; a stopped sandbox is started by API	Yes; apps scale to zero and start again on the next request, with autoscaling up to your replica limit
GPUs	No, CPU only	T4 to B200 and RTX PRO 6000, up to 8 per container, plus Inferentia2 and Trainium
Inbound access	HTTPS URLs for up to 50 ports, token-protected by default; raw TCP; dedicated IPv4 or IPv6	One endpoint per app on one port, open unless you set disable_auth = false; custom domains on Standard
Desktop and browser	Linux desktop streamed at 60fps with Chromium and computer use	No desktop or browser product
Coding agents	Seven agents preinstalled, with a prompt, steer and events API	None to run inside; an Agent Skill and a docs MCP server help your own agent deploy to Cerebrium
Keeping API keys out of the machine	Credentials proxy for model providers and any HTTPS API; switch a credential to another account when credits run out	Secrets are exposed to the app as environment variables
Regions	EU only (Germany, Finland, France)	US and EU generally available; Asia, South America, Canada and the Middle East on request
Compliance and transparency	SOC 2 is in progress. We're open about where we stand: live status and data retention policy.	SOC 2

Every feature Battleships tracks (43)

Feature	bochka	Cerebrium
Full VM (own kernel)	Yes	No
Docker inside	Yes	No
Nested virtualization	Yes	No
Root	Yes	Not published
Systemd	Yes	Not published
Desktop GUI	Yes	No
Browser	Yes	No
Computer-use API	Yes	No
SSH	Yes	No
Persistent disk	Yes	Yes
Pause/resume	Yes	No
Automatic snapshots	Yes	No
Fork/clone	Yes	No
Live fork (no pause)	Yes	No
Memory fork	No	No
Custom image (Docker or snapshot)	Yes	Yes
≥24 h sessions	Yes	Yes
Idle auto-stop	No	Yes
Wake on request	No	Yes
HTTPS preview URLs	Yes	Yes
Raw TCP inbound	Yes	Not published
Public IPv4	No	No
Egress allowlist	No	No
Secret proxy	Yes	No
Secret proxy for any API	Yes	No
Preinstalled agents	Yes	No
Agent harness API	Yes	No
MCP server	No	No
GPU	No	Yes
Windows	No	No
Arm64	No	No
Self-hosting	No	No
BYOC	No	No
Open source	No	No
Scoped API keys	Yes	Not published
Spend limits	Yes	Not published
Usage per sandbox (API)	Yes	Yes
Audit logs	No	Yes
SSO	No	No
SOC 2	No	Yes
HIPAA	No	Yes
EU data residency	Yes	Yes
Zero data retention	Yes	Not published

Performance

bochka26.69 runs/s

CerebriumNot measured

There is no comparable measurement: Cerebrium has no result on the StarSling hpc-sandbox-benchmarks CPU test, so we cannot say which is faster at CPU work. bochka did 26.69 runs per second on that test at 4 vCPU / 8 GB. Cerebrium's own speed claims are about GPU inference and cold starts, which this test does not cover.

Node.js build-and-test loop at 4 vCPU / 8 GB, higher is better. Measured with the open hpc-sandbox-benchmarks suite. Cerebrium has 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	bochka, a month	Cerebrium, 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	$20 · $20/mo	$53 · Hobby · with limits
One agent that never stops 4 vCPU / 8 GB, 50 GB disk, running 24/7 all month	$26 · $20/mo	$210 · Hobby · with limits
50 agents in parallel 50 machines of 4 vCPU / 8 GB at once, 8-hour sessions, 1,100 sessions a month, Docker inside	$317 · $20/mo × 3 seats	$2,557 · Hobby · with limits
Sandboxes for your users 2 vCPU / 4 GB, 30,000 sessions of 10 minutes a month, up to 300 at once	$200 · $100/mo × 2 seats	$715 · Hobby
1,000 short jobs at once 2 vCPU / 4 GB, 1,000 at once, 3 minutes each, 200,000 jobs a month	$2,000 · $2000/mo · with limits	$1,681 · Standard

bochka: Plans start at $20 a month, and the whole fee is usage credit. You pay per second while a sandbox runs, and a stopped sandbox costs nothing and keeps its disk.

Cerebrium: Cerebrium bills per second for each vCPU, GB of memory and GPU while a container runs, and does not bill the platform's cold-start time. The default tier can be interrupted; the protected tier costs 2x on all compute. Standard is $100 a month on top of compute, with no usage credit, and storage past the first 100 GB costs $0.05 per GB a month.

Limits in those scenarios

Cerebrium does not fit the coding agent, always-on and 50-agent scenarios, for the same reason: only /persistent-storage survives a stop, so an agent's installed packages reset, and the 50-agent fleet also needs Docker inside, which is not documented. An always-on agent is billed for every hour it runs, since a container cannot stop and keep its state for free. Cerebrium does not publish an egress price, so the scenarios count none. bochka's start limits are lower: at 1,000 short jobs at once, its $2,000 plan queues starts, while Cerebrium's Standard plan fits the burst.

Scenario	bochka	Cerebrium
One coding agent, workdays	None that bite	files survive a stop only in /persistent-storage: packages installed elsewhere reset egress price not published
One agent that never stops	None that bite	files survive a stop only in /persistent-storage: packages installed elsewhere reset egress price not published
50 agents in parallel	None that bite	missing Docker inside files survive a stop only in /persistent-storage: packages installed elsewhere reset egress price not published
Sandboxes for your users	None that bite	egress price not published
1,000 short jobs at once	max 2,400 starts/day (needs ~6,667) max 600 starts/hour (needs ~1,000) bursts above 90 starts/min are queued (a burst of 333 takes ~4 min to start)	egress price not published

Which one should you use?

Use Cerebrium when

you need GPUs to serve a model, voice or video app as an endpoint that scales with traffic, or when it must run in the US or Asia.

GPUs for serving models. For model inference, Cerebrium offers GPUs from T4 to B200, up to 8 in one container, billed per second.
Endpoints that scale to zero. Apps that wait for traffic can scale to zero, start on the next request, and grow to many replicas under load.
More regions. It runs in the US and the EU by default, with Asia, South America, Canada and the Middle East on request.

Use bochka when

your agents need a machine of their own: a shell, root, Docker, emulators and a browser, on a full VM whose whole disk is still there the next time it starts.

A whole machine to work inside. A coding agent installs tools, starts services and runs docker build as it goes. On Cerebrium you would have to wrap every step in an endpoint call, and anything needing Docker or /dev/kvm has no documented way to run.
The whole disk stays, and you can fork it. An agent that pip installs, clones repos and edits config outside one folder loses all of it on Cerebrium when the container scales down, so each run has to set itself up again from scratch.
Agents, desktop and keys built in. Cerebrium has no agent harness or desktop to start from, so you build both, and its secrets arrive as environment variables inside the container, where a prompt-injected agent can read and leak them.

Questions people ask

Is Cerebrium a sandbox for AI agents?

Not in the usual sense. Cerebrium is a serverless GPU cloud that runs your app as autoscaling containers behind endpoints. It has no SSH, no desktop and no agent harness, while a bochka sandbox is a full VM an agent works inside.

Is bochka cheaper than Cerebrium?

In four of the five scenarios on this page, yes, such as $26 against $210 a month for one agent that never stops. For 1,000 short jobs at once, Cerebrium's Standard plan fits the burst while bochka's plans queue the starts.

Can bochka run GPUs like Cerebrium?

No. bochka is CPU only. If your agent needs a GPU to serve or run a model, Cerebrium offers GPUs from T4 to B200.

What happens to files when a Cerebrium container stops?

Only files in the /persistent-storage volume are kept; the rest of the container resets, so packages installed elsewhere are gone. bochka saves the whole disk every minute and keeps it while the sandbox is stopped.

Can I run Docker inside Cerebrium?

Cerebrium builds your app from a Dockerfile, but running Docker inside the app is not documented. A bochka sandbox runs Docker, systemd and /dev/kvm like a normal Ubuntu server.

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.

Start a sandbox

Install the CLI, then run bochka new.

Product preview