boat vs Baseten

Baseten is an inference cloud: you deploy models on its GPUs, call hosted open models, and train. It launched Sandboxes for agents on November 5, 2026, free until November 16. boat gives each agent a full VM with root, Docker, a desktop and coding agents installed. boat saves the disk every minute while the sandbox runs, and once more at each stop.

Where boat stands out

A full VM with documented root and Docker

On boat every sandbox has its own kernel, with root via sudo, real systemd, Docker and /dev/kvm ready. Baseten model containers run as the non-root app user. Launch coverage says each Baseten sandbox gets its own microVM, but the sandbox docs do not document root, Docker or /dev/kvm inside it.

A kept disk that you can fork

boat saves the disk every minute while a sandbox runs, and once more at each stop. A stopped sandbox has no compute charge, and boat keeps its disk for at least 14 days; a saved snapshot has no scheduled expiry. You can fork a running sandbox from its last 1-minute snapshot, so one set-up machine becomes many. Writes from the last minute or so can be missing from the fork. Baseten's standby keeps more state (memory and processes), but its docs show no fork.

Coding agents and a desktop installed

boat ships seven coding-agent harnesses, including Claude Code, Codex and OpenCode, inside each sandbox, and you drive them from outside with a prompt, steer and events API. Every sandbox also has a 1920x1080, 60 fps desktop with Chrome and computer tools. On Baseten you set these up yourself: install the agent and a desktop in a custom sandbox image, or drive the sandbox from an agent outside through the CLI or the sandbox's MCP endpoint.

The companies and products

boat

boat.dev 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.

Baseten

Baseten is an inference cloud: it runs AI models on GPUs for production apps, sells hosted open models and training, and now runs agent sandboxes.

You package a model with Truss, its open-source format, or bring a Docker image, and Baseten runs it as replicas on Kubernetes that scale up with traffic and down to zero. It also sells Model APIs (open models priced per token) and training jobs on the same GPUs. Sandboxes, isolated environments where an agent runs commands and works with files, launched on November 5, 2026. They are free until November 16, 2026, and later pricing is not published. This page compares boat with both the CPU instances and the Sandboxes.

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. “5th cheapest of 365” means 4 of the other 364 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 workloadboat: price rankBaseten: price rank
Software factory
100 at once · 4 vCPU / 8 GB · 2,200 sessions of 8 h a month
Baseten model containers run as the non-root app user, and its sandbox docs do not document root, Docker or /dev/kvm. Test stacks that need Docker run on a boat VM, which has root and Docker on its own kernel.
5th cheapest of 365
$634 a month
234th of 365, with limits
$2,619 a month
Always-on AI employees
50 machines 24/7 · 2 vCPU / 4 GB
On Baseten, set an idle or date expiry, or mount a sandbox volume, so files outlive the 7-day default. On boat, a stopped disk stays for at least 14 days, and a saved snapshot has no scheduled expiry.
6th cheapest of 361
$657 a month
278th of 361, with limits
$3,258 a month
Code interpreter bursts
100 at once · 1 vCPU / 2 GB · 200,000 sessions of 1 min a month
Baseten CPU instances bill per minute with no plan fee, but each burst also pays cold start and, by default, 15 idle minutes. Baseten Sandboxes bill only storage on standby, with no published price after November 16, 2026. boat starts at $20 a month.
189th of 336, with limits
$2,000 a month
313th of 336, with limits
$1,856 a month
Sandboxes for your users
300 at once · 2 vCPU / 4 GB · 30,000 sessions of 23 min a month
Baseten Sandboxes run up to 10 at once per organization by default, and support raises the limit. boat runs 300 at once on the $100 plan.
3rd cheapest of 349
$209 a month
273rd of 349, with limits
$1,868 a month
RL rollouts
1,000 at once · 2 vCPU / 4 GB · 1,000,000 sessions of 3 min a month
39th of 349, with limits
$2,000 a month
315th of 349, with limits
$31,140 a month
Agent evals
250 at once · 4 vCPU / 8 GB · 15,000 sessions of 20 min a month
10th cheapest of 359
$200 a month
297th of 359, with limits
$1,609 a month
CI runners
40 at once · 8 vCPU / 16 GB · 6,000 sessions of 12 min a month
A CI job needs Docker and root for most builds. boat gives root and Docker on every VM. Baseten documents neither for its model containers or its sandboxes.
2nd cheapest of 374
$86 a month
323rd of 374, with limits
$1,119 a month
24/7 dev boxes
1 machine 24/7 · 4 vCPU / 8 GB
19th cheapest of 371
$26 a month
186th of 371, with limits
$130 a month
Computer-use agents
50 at once · 4 vCPU / 8 GB · 20,000 sessions of 15 min a month
Baseten documents no desktop or computer-use API. You set it up: add a desktop and a VNC web client to a custom sandbox image, and open it through a preview URL. boat streams a 1920x1080, 60 fps desktop with Chrome and computer tools.
7th cheapest of 367
$200 a month
314th of 367, with limits
$1,799 a month
Browser-use agents
50 at once · 2 vCPU / 4 GB · 30,000 sessions of 10 min a month
7th cheapest of 356
$200 a month
331st of 356, with limits
$1,194 a month
Scraping & bots (startup scale)
50 at once · 1 vCPU / 2 GB · 100,000 sessions of 3 min a month
6th cheapest of 343
$90 a month
275th of 343, with limits
$183 a month
Mass scraping & bots (undetectable)
1,000 at once · 1 vCPU / 2 GB · 3,000,000 sessions of 2 min a month
73rd of 327, with limits
$1,800 a month
128th of 327, with limits
$3,654 a month
Parallel red teaming
200 at once · 2 vCPU / 4 GB · 5,000 sessions of 30 min a month
3rd cheapest of 351
$200 a month
275th of 351, with limits
$381 a month
GPU eval jobs
10 at once · 8 vCPU / 32 GB · 500 sessions of 30 min a month
Baseten fits: it rents Nvidia GPUs from T4 to B200, up to 8 per instance. boat has no GPU on any size.
Not ranked: misses a requirement25th cheapest of 102
$1,625 a month
macOS agents
5 at once · 4 vCPU / 8 GB · 1,000 sessions of 20 min a month
Not ranked: misses a requirementNot ranked: misses a requirement

Key features

FeatureboatBaseten
Kind of productSandboxes for AI agents: VMs with agents, desktop and hosting readyInference cloud for serving and training models; agent Sandboxes launched November 5, 2026, CLI labeled pre-release
IsolationEvery sandbox is a full VM with its own kernelSandboxes: a microVM per sandbox with its own kernel, per launch coverage (Baseten's docs do not name it). Model deployments: containers
Root, Docker and nested virtualizationRoot via sudo, real systemd, Docker and /dev/kvm on every sandboxModel deployments run as the non-root app user; the sandbox docs do not document root, Docker or /dev/kvm
Getting a sandboxSelf-serve on every plan, 100 to 2,000 at once by plan; the 7-day trial needs a card and runs 2 at onceNo access request documented; 10 per organization at once by default, more through support
Sandbox size4 sizes, 2 vCPU / 4 GB to 16 vCPU / 32 GB; disk fixed per size, 51 to 251 GB1 to 32 vCPU, 1 to 8 GiB per vCPU; size fixed after creation
Files after a stopDisk saved every minute and at each stop; kept at least 14 days, saved snapshots longer; no compute charge while stopped; memory and processes are not keptStandby keeps files, memory and running processes and bills only storage; sandboxes expire 7 days after creation by default (configurable); sandbox volumes keep files after deletion
ForkFork a running sandbox into many, from its last 1-minute snapshotNo fork documented; you start a new sandbox from the image, and a volume mounts on one sandbox at a time
Scale to zero and wake on requestNo wake on request; you resume a stopped sandbox through the API, and a resume is like a rebootDeployments scale to zero and start on the next request; standby sandboxes resume with memory and processes on the next request
GPUsNone on any sizeNvidia T4 up to B200, 1 to 8 per instance, multi-node over InfiniBand
Desktop and computer use1920x1080, 60 fps desktop and Chrome, computer tools for agents, embeddable streamNone documented; you set it up: add a desktop and a VNC web client to a custom sandbox image and open it through a preview URL
Coding agentsSeven harnesses installed, including Claude Code, Codex and OpenCode, with a prompt, steer and events APIYou set it up: install the agent in a custom sandbox image, or drive the sandbox from outside through the CLI or its MCP endpoint; Baseten Switch routes Claude Code or Codex to Model APIs
Reaching the sandboxUp to 50 HTTPS URLs per sandbox, token-gated by default; SSH; public IPv6 and outbound IPv4; raw ports (over IPv6 only on sandboxes with no IPv4 of their own)HTTPS API endpoint per deployment; sandbox previews give public or private HTTPS URLs per port, plus a /port/ route with a token
Outbound networkOpen internet on every plan; no provider-side allowlistDeployments: TCP ports 443 and 80 by default; per-deployment IP and domain allowlists once enabled for your organization. Not documented for sandboxes
Keeping keys out of the sandboxOptional AI credentials proxy, $1 a day while on plus $1 per 1,000 requests; real keys of supported providers stay out; other secrets still enterSecrets are passed into the model as arguments or files under /secrets
Regions and hostingGermany, Finland and France; you cannot pick a region; no US or Asia regionUS and EU for GPU deployments; self-hosted or hybrid in your own cloud account on Enterprise
Compliance and transparencySOC 2 is in progress. We're open about where we stand: live status and data retention policy.SOC 2
Every feature Battleships tracks (43)
FeatureboatBaseten
Full VM (own kernel)YesNo documented way
Docker insideYesNo documented way
Nested virtualizationYesNo documented way
RootYesNot published
SystemdYesNot published
Desktop GUIYesNo documented way
BrowserYesNo documented way
Computer-use APIYesNo documented way
SSHYesYes
Persistent diskYesNo documented way
Pause/resumeYesNo documented way
Automatic snapshotsYesNo documented way
Fork/cloneYesNo documented way
Live fork (no pause)YesNo documented way
Memory forkNo documented wayNo documented way
Custom image (Docker or snapshot)YesYes
≥24 h sessionsYesYes
Idle auto-stopNo documented wayYes
Wake on requestNo documented wayYes
HTTPS preview URLsYesYes
Raw TCP inboundYesNot published
Public IPv4No documented wayNo documented way
Egress allowlistNo documented wayYes
Secret proxyYesNo documented way
Secret proxy for any APIYesNo documented way
Preinstalled agentsYesNo documented way
Agent harness APIYesNo documented way
MCP serverNo documented wayYes
GPUNo documented wayYes
WindowsNo documented wayNo documented way
Arm64No documented wayNo documented way
Self-hostingNo documented wayYes
BYOCNo documented wayYes
Open sourceNo documented wayNo documented way
Scoped API keysYesNot published
Spend limitsYesNo documented way
Usage per sandbox (API)YesYes
Audit logsNo documented wayNot published
SSONo documented wayYes
SOC 2No documented wayYes
HIPAANo documented wayYes
EU data residencyNo documented wayYes
Zero data retentionYesNot published

Performance

boat28.06 runs/s
BasetenNot measured

Baseten is not in the public benchmarks we use, so we cannot say which is faster. In StarSling's run of 3 October 2026 (run 37073680030), boat did 28.06 runs/s on the V8 Web Tooling Benchmark at 4 vCPU / 8 GB, second of 15. That boat sandbox ran on an AMD Ryzen 9 9950X host. boat's fleet also has Ryzen 7 9700X hosts, and boat does not promise a CPU model.

V8 Web Tooling Benchmark at a 4 vCPU / 8 GB target, runs per second, higher is better. Source: StarSling's run of 3 October 2026 (run 37073680030), median of 5 sandboxes. Providers ran on different hardware in different regions. StarSling gives a shared rank when the run cannot separate two providers. Every boat sandbox in this run ran on an AMD Ryzen 9 9950X host. boat also places some sandboxes on AMD Ryzen 7 9700X hosts, which this run did not measure. Baseten is not in the public benchmarks we use, so we cannot say which is faster.

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. Start limits, per-start fees and plan minimums are in the price, boat's included.

Scenarioboat, a monthBaseten, 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$34 · Basic · with limits
One agent that never stops
4 vCPU / 8 GB, 50 GB disk, running 24/7 all month
$26 · $20/mo$136 · Basic · 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$1,701 · Basic · 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$1,246 · Basic · with limits
1,000 short jobs at once
2 vCPU / 4 GB, 1,000 at once, 3 minutes each, 200,000 jobs a month
$8,800 · $2000/mo · with limits$6,228 · Basic · with limits

boat: Plans cost $20, $100, $500 or $2,000 a month. The whole plan price comes back as VM time for that month, and unused time expires at the end of the period. A 4 vCPU / 8 GB sandbox costs $0.036 an hour, billed per second while it runs. A stopped sandbox has no compute charge, but extra saved snapshots and the AI credentials proxy still charge. Above the included starts (200 a day on $20), each extra start costs $0.03 to $0.07.

Baseten: Baseten Sandboxes are free until November 16, 2026, and pricing after that is not published, so the scenarios use its CPU instances. These bill per minute with partial minutes rounded up, from image build and cold start until the replica stops. Baseten has no 4 vCPU / 8 GiB CPU instance, so the scenarios use 4 vCPU / 16 GiB at $0.00346 a minute ($0.2076 an hour). There is no plan fee on Basic.

Limits in those scenarios

Baseten's CPU instances are model containers with no persistent disk, so files are lost when a replica stops or scales to zero, and the table marks them as not fitting. Baseten Sandboxes, its agent product, keep files on standby and can mount volumes, but have no published price after November 16, 2026, so the table does not price them. They run 10 at once per organization by default, and they expire 7 days after creation unless you set another policy. A Baseten monthly budget never stops dedicated deployments; enforcement covers Model APIs only. On boat the limit is a sandbox count by plan plus starts per minute, hour and day. boat never queues a start: a start above 3 times the per-minute limit fails with 429. The 1,000-job burst needs about 6,667 starts a day, against 2,400 included on the $2,000 plan, so it does not fit on boat either.

ScenarioboatBaseten
One coding agent, workdaysNone for this scenario
  • adjacent product, not a like-for-like sandbox
  • missing persistent disk
  • different kind of product: app platform
  • no persistent disk: files lost when it stops or sleeps
One agent that never stopsNone for this scenario
  • adjacent product, not a like-for-like sandbox
  • missing persistent disk
  • different kind of product: app platform
  • no persistent disk: files lost when it stops or sleeps
50 agents in parallelNone for this scenario
  • adjacent product, not a like-for-like sandbox
  • missing persistent disk, Docker inside
  • different kind of product: app platform
  • no persistent disk: files lost when it stops or sleeps
Sandboxes for your usersNone for this scenario
  • adjacent product, not a like-for-like sandbox
  • different kind of product: app platform
  • disk beyond 0 GiB: price unknown
  • egress price not published
1,000 short jobs at once
  • 2,400 starts/day included, paid above that (needs ~6,667)
  • 600 starts/hour included, paid above that (needs ~1,000)
  • starts above 90/min are paid, and starts above 270/min fail with 429
  • adjacent product, not a like-for-like sandbox
  • different kind of product: app platform
  • disk beyond 0 GiB: price unknown
  • egress price not published
boat's own limits
  • Plans start at $20 a month. The plan price comes back as VM time for that month. Unused time does not carry over. Source
  • Starts (create, fork, resume): the $20 plan includes 12 a minute, 60 an hour and 200 a day. The $2,000 plan includes 90 a minute, 600 an hour and 2,400 a day. Extra starts cost $0.03 to $0.07. A start fails with 429 above 3 times the included starts per minute, or 6 times per hour or per day. Source
  • A stopped sandbox has no compute charge. boat keeps its disk for at least 14 days. A saved snapshot keeps it longer; extra saved snapshots and the AI credentials proxy are paid. Source
  • Auto-stop defaults to 1 hour. You can turn it off on a paid plan. The trial allows 2 hours at most. Source
  • boat saves the disk every minute and at each stop. Memory, running programs and /tmp are not saved: a resume is like a reboot. A machine failure loses writes after the last good snapshot. Source
  • Machines run in Germany, Finland and France only. Snapshot copies can be stored outside the EU. Source
  • Public IPv6 and outbound IPv4. No dedicated IPv4. The IP changes at each resume. Source
  • Hosts are AMD Ryzen 7 9700X and Ryzen 9 9950X. vCPUs are shared. No GPU. You cannot boot your own Docker image. Source

Which one should you use?

Use Baseten when

you serve or train models on GPUs, want deployments that scale to zero and wake on request, want sandboxes that keep memory on standby or go above 16 vCPU, or must run in the US or in your own cloud account.

  • GPUs for serving and training models. Useful if your product serves or trains models: Baseten runs Nvidia GPUs from T4 to B200, up to 8 per instance.
  • Scale to zero, wake on request, memory kept on standby. Useful for spiky traffic and idle agents: deployments scale to zero, and sandboxes resume with memory and processes on the next request.
  • Bigger sandboxes, US region, your own cloud. Useful if you need up to 32 vCPU per sandbox, the US, your own cloud account, or per-deployment egress allowlists.

Use boat when

you run agents that build and test software: full VMs with documented root and Docker, coding agents and a desktop installed, a disk saved every minute, and forks of a running sandbox.

  • A full VM with documented root and Docker. Coding agents install packages, start databases with Docker and run emulators. On boat this works on day one, with no image to build.
  • A kept disk that you can fork. Agents branch their work. On Baseten you start each new sandbox from the image and copy state in, and a volume mounts on one sandbox at a time.
  • Coding agents and a desktop installed. Without these you build the image, the agent loop, the browser and the desktop stream yourself. boat's optional credentials proxy ($1 a day while on, plus $1 per 1,000 requests) also keeps real AI keys of supported providers off the machine, where Baseten passes secrets into the container.

Questions people ask

Is boat cheaper than Baseten?

It depends on the product. Baseten Sandboxes are free until November 16, 2026, so until then they cost less than boat; later pricing is not published. Against Baseten's CPU instances at list price, boat costs less in 4 of our 5 scenarios: one coding agent on workdays is $20 a month on boat and $34 on Baseten, one agent running all month is $26 versus $136, 50 agents in parallel are $317 versus $1,701, and 300 sandboxes for your users are $200 versus $1,246. Those CPU instances keep no files after a stop and have 16 GiB of RAM against boat's 8 GB. For 1,000 short jobs at once neither fits: Baseten comes to $6,228 and boat to $8,800 with paid extra starts.

Does Baseten have sandboxes for AI agents?

Yes. Baseten Sandboxes launched on November 5, 2026 and are free until November 16, 2026; later pricing is not published, and the CLI is labeled pre-release. A sandbox runs commands and file operations from an image, with 1 to 32 vCPU. On standby it keeps files, memory and running processes and bills only storage. It expires 7 days after creation by default, which you can change, and volumes keep files after deletion. An organization runs up to 10 at once by default.

Does boat have GPUs like Baseten?

No. boat runs Linux VMs on AMD Ryzen hosts and has no GPU on any size. If your product serves or trains models, run that part on Baseten and let your agents work on boat.

Can I run Docker inside Baseten?

Baseten does not document Docker inside its model containers or its sandboxes, and its model containers run as the non-root app user. You can bring your own Docker image for Baseten to run as a model server, or build a custom sandbox image from a Dockerfile. Every boat sandbox has root and Docker with Compose on its own kernel.

Where does my data run?

boat sandboxes run on machines in Germany, Finland and France, and you cannot pick a region. Snapshot copies can be stored outside the EU. Baseten lets GPU deployments pick a US or EU region, its shared CPU instances cannot pick one, and Enterprise customers can self-host in their own cloud.

Is boat SOC 2 compliant?

SOC 2 is in progress, so there is no SOC 2 report today, and no HIPAA BAA or SAML SSO is documented. Baseten says it is SOC 2 Type II certified and HIPAA compliant, and offers SAML SSO on Enterprise. Live status is at status.ascii.dev and our data retention policy at docs.boat.dev/data-retention.

Start a sandbox

Install the CLI, then run boat new.

7-day free trial

More comparisons