CodeSandbox SDK vs exe.dev

CodeSandbox SDK starts Firecracker VMs that hibernate with their memory when idle and wake on the next HTTP request, driven from a TypeScript SDK. exe.dev rents persistent KVM VMs you reach over SSH, from a reserved pool or billed per second, with coding agents preinstalled and nine regions to pick from.

The companies and products

CodeSandbox SDK

CodeSandbox SDK is an API from CodeSandbox, now a Together AI company, that starts sandboxed development VMs on the same infrastructure as the CodeSandbox editor.

Each sandbox is a Firecracker microVM built from a template or any Dockerfile. When it goes idle it hibernates with a memory snapshot, and it resumes or forks from that snapshot in seconds. You drive it from a TypeScript SDK through terminals, commands, files and HTTPS hosts. Together's docs now call it the legacy offering and point new users to its successor, the together-sandbox SDK.

Self-hosting
Not offered

exe.dev

exe.dev, made by Bold Software, Inc., rents Linux VMs for developers and AI agents that you reach over SSH and HTTPS.

You reserve a pool of vCPUs and memory, and your VMs share it; you can also start standalone VMs billed by the second. Each VM is KVM-isolated (they use Cloud Hypervisor), boots from a container image in about two seconds, and has root, systemd and Docker. It comes with Claude Code, Codex, Pi and exe.dev's own Shelley agent, and is aimed at developers who want a persistent dev box or a sandbox for their agents.

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. “46th cheapest of 345” means 45 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 workloadCodeSandbox SDK: price rankexe.dev: price rankboat: price rank
Software factory
100 at once · 4 vCPU / 8 GB · 2,200 sessions of 8 h a month
178th of 367, with limits
$1,939 a month
253rd of 367, with limits
$3,701 a month
3rd cheapest of 367
$634 a month
Always-on AI employees
50 machines 24/7 · 2 vCPU / 4 GB
exe.dev fits an always-on assistant fully with a VM that runs until you stop it; CodeSandbox SDK gets costly running all month, and boat adds a desktop and an agent API.
215th of 366, with limits
$2,478 a month
15th cheapest of 366
$3,833 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
CodeSandbox SDK fits short code runs well, since VMs hibernate between runs and wake fast; exe.dev is built for long-lived VMs, not many tiny runs.
46th cheapest of 345
$642 a month
235th of 345, with limits
$175 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 fits 300 user sandboxes at once on its self-serve plans (250 on CodeSandbox SDK Scale, 50 on exe.dev Personal); boat does.
140th of 361, with limits
$1,707 a month
257th of 361, with limits
$1,226 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
CodeSandbox SDK's memory forks let rollouts branch from one running state, while exe.dev's cp copies only the disk.
348th of 360, with limits
$3,861 a month
265th of 360, with limits
$5,250 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
CodeSandbox SDK fits agent evals, with forks from one warm memory state; exe.dev's 50 VM cap on Personal gets in the way, and boat fits at higher concurrency.
64th cheapest of 364
$1,335 a month
283rd of 364, with limits
$1,050 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
CodeSandbox SDK bills only while jobs run, which suits CI better than an exe.dev pool that bills whether jobs run or not; boat adds /dev/kvm for emulator builds.
92nd cheapest of 377
$860 a month
288th of 377, with limits
$504 a month
4th cheapest of 377
$86 a month
24/7 dev boxes
1 machine 24/7 · 4 vCPU / 8 GB
exe.dev makes a good 24/7 dev box with SSH and a pool billed flat; CodeSandbox SDK has no SSH and bills credits for every running minute, and boat fits too with a desktop in the browser.
310th of 368, with limits
$240 a month
25th cheapest of 368
$35 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
Neither CodeSandbox SDK nor exe.dev has a desktop for computer-use agents; boat streams one at 60fps.
206th of 369, with limits
$691 a month
240th of 369, with limits
$1,055 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
251st of 367, with limits
$443 a month
130th of 367, with limits
$530 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
281st of 346, with limits
$508 a month
124th of 346, with limits
$278 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
193rd of 338, with limits
$7,390 a month
166th of 338, with limits
$5,490 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
240th of 360, with limits
$326 a month
60th of 360, with limits
$263 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 itCan't run itCan't run it
macOS agents
5 at once · 4 vCPU / 8 GB · 1,000 sessions of 20 min a month
Can't run itCan't run itCan't run it

Key features

FeatureCodeSandbox SDKexe.devboat
IsolationFirecracker microVM with its own kernelKVM VM with Cloud Hypervisor; you can't choose the kernelFull VM with its own kernel on dedicated AMD Ryzen 9 9950X hosts
Pause and resumeHibernates with a memory snapshot; resumes with processes still runningDisk persists and is streamed to a second server; no snapshotsDisk is saved every minute and on stop; resume boots the machine again
ForkFork a hibernated sandbox from its snapshot, or a running one sharing its memorycp copies a VM's disk and config into a new VMFork a running sandbox with its disk; named snapshots as templates
Wake on requestYes; an HTTP request to a host wakes a hibernated VMNo; VMs run until you stop themNo; a stopped sandbox is started by API
Shell accessTerminals, commands and files through the SDK; no SSHSSH, brokered through exe.devSSH, plus exec and files over the API
Custom imagesBuilt from a template or any DockerfileBoot from your own Docker imageTemplates from named snapshots; no Dockerfile or OCI images
Docker and rootDocker and rootDocker, systemd and rootDocker, systemd and root
Nested virtualizationNot documentedEnterprise onlyYes, /dev/kvm on every plan
Coding agentsRun your own agent through the command APIClaude Code, Codex, Pi and Shelley preinstalled; only Shelley takes prompts by APIClaude Code, Codex and five more preinstalled, with a prompt, steer and events API
Desktop and browserNo desktop; a headless Chromium templateHeadless; no desktop or screen streamingLinux desktop streamed at 60fps, Chromium and computer-use tools
Inbound accessHTTPS host per port, public by default; no raw TCP or public IPHTTPS proxy for ports 3000 to 9999 and SSH; no public IP or raw TCPHTTPS URLs for up to 50 ports, token-protected by default; raw TCP; dedicated IPv6 or IPv4
Keeping API keys out of the sandboxSecrets passed in as environment variablesIntegrations inject secrets at the network edge for any HTTP hostCredentials proxy for model providers and any HTTPS API; switch accounts when credits run out
RegionsNo region choice; a country hint steers scheduling but is not guaranteed9 regions in the US, Europe, Asia and AustraliaEU only: Germany, Finland, France
Biggest machineUp to 16 cores / 32 GB on Scale, 64 cores / 128 GB on EnterprisePersonal pools up to 16 vCPUUp to 16 vCPU / 32 GB on the $100 plan and above
Every feature Battleships tracks (43)
FeatureCodeSandbox SDKexe.devboat
Full VM (own kernel)YesYesYes
Docker insideYesYesYes
Nested virtualizationNoYesYes
RootYesYesYes
SystemdNot publishedYesYes
Desktop GUINoNoYes
BrowserNoYesYes
Computer-use APINoNoYes
SSHNoYesYes
Persistent diskYesYesYes
Pause/resumeYesYesYes
Automatic snapshotsYesNoYes
Fork/cloneYesYesYes
Live fork (no pause)YesYesYes
Memory forkYesNoNo
Custom image (Docker or snapshot)YesYesYes
≥24 h sessionsYesYesYes
Idle auto-stopYesNoNo
Wake on requestYesNoNo
HTTPS preview URLsYesYesYes
Raw TCP inboundNoNoYes
Public IPv4NoNoNo
Egress allowlistNoNoNo
Secret proxyNoYesYes
Secret proxy for any APINoYesYes
Preinstalled agentsNoYesYes
Agent harness APINoYesYes
MCP serverNoYesNo
GPUNoNoNo
WindowsNoNoNo
Arm64NoNoNo
Self-hostingNoNoNo
BYOCNoYesNo
Open sourceNoNoNo
Scoped API keysNot publishedYesYes
Spend limitsNot publishedNot publishedYes
Usage per sandbox (API)NoYesYes
Audit logsNot publishedNot publishedNo
SSOYesYesNo
SOC 2YesNoNo
HIPAANoNoNo
EU data residencyNoYesYes
Zero data retentionNot publishedNot publishedYes

Performance

CodeSandbox SDKNot measured
exe.devNot measured
boat26.69 runs/s

Neither CodeSandbox SDK nor exe.dev is in the hpc-sandbox-benchmarks run, so there is no comparable CPU measurement for either. boat did 26.69 runs per second on that Node.js loop at 4 vCPU / 8 GB, on dedicated AMD Ryzen 9 9950X hosts. We don't guess how the other two would score.

Node.js build-and-test loop at 4 vCPU / 8 GB, higher is better. Measured with the open hpc-sandbox-benchmarks suite. CodeSandbox SDK and exe.dev 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.

ScenarioCodeSandbox SDK, a monthexe.dev, a monthboat, 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
$193 · Scale$35 · Personal$20 · $20/mo
One agent that never stops
4 vCPU / 8 GB, 50 GB disk, running 24/7 all month
$341 · Scale$35 · Personal$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
$2,500 · Scale$1,848 · Personal$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
$815 · Scale · with limits$530 · Personal · with limits$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
$1,632 · Scale · with limits$1,050 · Personal · with limits$2,000 · $2000/mo · with limits

CodeSandbox SDK: You buy VM credits and each size burns a fixed number per hour, billed per started minute while the VM runs, including idle time before it hibernates. The Scale plan adds a $170 monthly platform fee that is not usage credit, and a hibernated VM costs nothing.

exe.dev: Pools are reserved capacity your VMs share, billed the same whether they are busy or idle: Personal is $35 a month for 4 vCPU / 8 GiB. Standalone VMs bill per second at $0.105 per 2 vCPU / 4 GiB per hour.

exe.dev: Plans start at $20 a month, and the whole fee is usage credit. Sandboxes bill per second while they run, and a stopped sandbox costs nothing and keeps its disk.

Limits in those scenarios

CodeSandbox SDK's Scale plan allows 250 VMs at once and exe.dev Personal allows 50, so neither fits the 300 user sandboxes or the 1,000 short jobs at once. CodeSandbox SDK gives each VM a 20 GB disk and does not publish the price for more disk or for egress. An exe.dev pool bills whether it is used or not. boat fits 300 at once, but its start limits queue a burst of 1,000 short jobs even on the $2,000 plan.

ScenarioCodeSandbox SDKexe.devboat
One coding agent, workdays
  • disk beyond 20 GiB: price unknown
  • egress price not published
  • flat pool: the tier is billed whether used or not
None that bite
One agent that never stops
  • disk beyond 20 GiB: price unknown
  • egress price not published
  • flat pool: the tier is billed whether used or not
None that bite
50 agents in parallel
  • egress price not published
None that biteNone that bite
Sandboxes for your users
  • Scale: max 250 concurrent
  • egress price not published
  • Personal: max 50 concurrent
None that bite
1,000 short jobs at once
  • Scale: max 250 concurrent
  • egress price not published
  • Personal: max 50 concurrent
  • 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)

Which one should you use?

Use CodeSandbox SDK when

your sandboxes sleep between visits and should wake on the next HTTP request with their memory intact, and you are fine building on what Together now calls its legacy SDK.

  • Memory hibernation and wake on request. Good for previews and code runs that sleep between visits: an HTTP request wakes the VM with its processes still running.
  • Forks that share memory. Useful for branching many rollouts or test variants from one warm running state instead of booting each one fresh.
  • Pay only while a VM runs. Credits burn per started minute while a VM runs, and a hibernated VM costs nothing, which suits bursty work.

Use exe.dev when

you want a persistent VM you SSH into, billed as a flat pool, in a region near users outside Europe.

  • A persistent box you SSH into. Good for developers who want a VM that runs until they stop it, with SSH, systemd and a flat monthly pool.
  • Regions around the world. Useful if your users are far from Europe: exe.dev runs in nine regions across the US, Europe, Asia and Australia.
  • Agents and team controls included. Claude Code, Codex and Pi come preinstalled, and the Work plan covers a whole team with SSO and roles.

Where boat fits

If you are weighing CodeSandbox SDK against exe.dev to run coding agents, look at boat too. It gives each agent a full VM with SSH, /dev/kvm, a 60fps desktop and its agent already set up behind one API, with state saved every minute and no cap on how long a paid sandbox runs.

A screen your agents can use

Neither CodeSandbox SDK nor exe.dev has a desktop, so a computer-use agent has nothing to look at. On both you build and stream a desktop yourself; boat streams one at 60fps with Chromium.

Drive agents inside from outside

CodeSandbox SDK leaves the agent harness to you, and exe.dev only takes API prompts for its own Shelley agent. boat lets you prompt, steer and stream events from Claude Code, Codex and others through one API.

Emulators and nested VMs on every plan

Android emulator builds and VM-based tests need /dev/kvm. CodeSandbox SDK does not document it, and exe.dev only offers it on Enterprise; boat has it on every plan.

Use boat when your agents do real work for hours: builds, Docker, emulators, a desktop and a browser, with state saved every minute and one API to drive agents inside from outside. Compare it one to one: boat vs CodeSandbox SDK and boat vs exe.dev.

Questions people ask

Is CodeSandbox SDK cheaper than exe.dev?

Not in our scenarios. One coding agent on workdays comes to $193 a month on CodeSandbox SDK against $35 on exe.dev, and 50 agents in parallel to $2,500 against $1,848. boat is lower than both in four of the five scenarios.

Which is faster, CodeSandbox SDK or exe.dev?

Neither has a comparable CPU measurement, since neither is in the hpc-sandbox-benchmarks run. boat scored 26.69 runs per second on that test at 4 vCPU / 8 GB.

Can I SSH into CodeSandbox SDK like exe.dev?

No. CodeSandbox SDK connects through its SDK, Node and browser clients, with terminals, commands and files. exe.dev and boat both give you SSH.

Can CodeSandbox SDK and exe.dev run Docker?

Yes, both run Docker with root inside the VM. For Android emulators or nested VMs, exe.dev needs Enterprise and CodeSandbox SDK does not document it, while boat has /dev/kvm on every plan.

Is CodeSandbox SDK still maintained?

CodeSandbox is now part of Together AI. Together's docs call the CodeSandbox-based SDK deprecated and name together-sandbox as its successor, while self-serve sign-up still runs through CodeSandbox.

Do CodeSandbox SDK or exe.dev have a desktop for computer use?

No. Both are headless. boat streams a 60fps desktop with Chromium and has APIs for computer use and browser control.

Try boat

Install the CLI, then run boat new.

7-day free trial

More comparisons