GPUBank vs JarvisLabs
JarvisLabs is a bootstrapped GPU cloud from New Delhi that rents 1 to 8 NVIDIA GPUs per instance, as managed containers or root VMs, from two regions in India. Its on-demand H100 SXM costs $2.69 per GPU-hour today and rises to $3.49 on 5 October 2026; GPUBank charges $1.85. JarvisLabs adds serverless inference and InfiniBand clusters; GPUBank bills by the second, includes local NVMe in the GPU rate, and runs in Europe, the US and Singapore.
GPUBank
Tallinn, Estonia
H100 SXM
$1.85 / h
- vCPU, RAM, NVMe, IPv4 and traffic included
- Per-second billing after the health check
- Top up in Bitcoin, USDT and 12+ other coins
JarvisLabs
New Delhi, India
H100 SXM
$2.69 / h
- Serverless inference, Managed Endpoints and token-priced Model APIs
- Self-serve H200 InfiniBand clusters, reserved clusters of 128 to 1,024+ GPUs
- A30 at $0.41 and L4 at $0.44 per GPU-hour, models GPUBank does not list
JarvisLabs
GPU cloud
28 points compared24 of their pages readprices last checked on 29 September 2026page updated 2 November 2026
The verdict
Serverless, clusters and small GPUs at JarvisLabs. Lower prices and four regions outside India at GPUBank.
JarvisLabs suits teams in India that want per-minute containers, INR invoices with GST, serverless endpoints and a self-serve H200 InfiniBand cluster. For the three GPUs both of us rent, GPUBank costs 31 to 39 % less per GPU-hour, and the gap widens when JarvisLabs raises its H100, H200 and RTX PRO 6000 rates on 5 October 2026. JarvisLabs bills disk on top of the GPU rate at $0.10 per GB-month and deletes instances and data at once if the balance goes negative. GPUBank bills by the second, charges $0.03 per GB-month for a stopped disk and takes crypto instead of cards.
01The real cost
The same GPUs, side by side, per hour.
Every GPU model both of us rent, at each provider's own published price for one card, in US dollars, excluding VAT. Prices last checked on 29 September 2026.
| GPU | JarvisLabs per hour | GPUBank per hour | Difference | 100 h at JarvisLabs | 100 h at GPUBank |
|---|---|---|---|---|---|
| $3.99 | $2.49 | −38 % | $399.00 | $249.00 | |
| $2.69 | $1.85 | −31 % | $269.00 | $185.00 | |
| $1.89 | $1.15 | −39 % | $189.00 | $115.00 |
JarvisLabs prices exclude tax and storage. On 5 October 2026 its on-demand H100 SXM rises to $3.49, H200 to $4.59 and RTX PRO 6000 to $2.19 per GPU-hour.
A fine-tune, 40 h on an H100 SXM
JarvisLabs
$107.60
GPUBank
$74.00
Saves $33.60 at GPUBank (−31 %)
A long-context job, 40 h on an H200
JarvisLabs
$159.60
GPUBank
$99.60
Saves $60.00 at GPUBank (−38 %)
A 70B inference box, 200 h on an RTX PRO 6000
JarvisLabs
$378.00
GPUBank
$230.00
Saves $148.00 at GPUBank (−39 %)
02Their catalogue
Every GPU JarvisLabs lists.
9 GPU models listed by JarvisLabs, each at its lowest public price per GPU-hour. 4 of them are also rented at GPUBank (in blue).
B300reservedNo price
B200180 GB · reservedNo price
H200141 GB · on demand$3.99
H100 SXM80 GB · on demand$2.69- A100 40GB40 GB · on demand$0.89
- A100 80GB80 GB · on demand$1.49
L424 GB · on demand$0.44
A3024 GB · on demand$0.41
RTX PRO 600096 GB · on demand$1.89At GPUBank
vCPU, RAM, NVMe and traffic included. All prices
03What the price includes
What appears on the bill besides the GPU.
The same H100 SXM session of 2 h 14 m 03 s, line by line, as each provider describes its billing.
JarvisLabs
2 h 14 m 03 s on H100 SXM
- GPU · $2.69 / h
- $6.01
- vCPU and RAM
- Included
- Storage
- $0.10 per GB-month
- Egress
- Free in-region
- Public IPv4
- Public IP included; static IP $5 per month
- Billing granularity
- Per minute
- Minimum charge
- None, no minimum rental period
GPUBank
2 h 14 m 03 s on H100 SXM
- GPU · $1.85 / h
- $4.13
- vCPU and RAM
- Included
- Storage
- Local NVMe included; persistent disk $0.03 per GB-month
- Egress
- Included
- Public IPv4
- Included
- Billing granularity
- Per second, from the health check
- Minimum charge
- None
While stopped
What keeps running on the bill.
JarvisLabs
Paused instance pays storage only.
A negative balance deletes instances and data at once, with no grace period.
GPUBank
Only the persistent disk, $0.03 per GB-month. Compute, IPv4 and scratch cost nothing.
04Where the GPUs run
1 country against 4, and where they are.
Countries where JarvisLabs says its GPUs run, from its own region list or documentation, against GPUBank's four regions. A pin marks a country, not the exact site.
JarvisLabs1
India
GPUBank4
Paris
Frankfurt
Ashburn
Singapore
JarvisLabs · 1 countryGPUBank · 4 regions
05Paying and signing up
How each one takes your money.
JarvisLabs
Payments and sign-up
Payment methods
Minimum top-up $5 in the docs, $10 on the home page
To start
- 1Email and password
- 2First and last name
- 3Phone number
Billing modelPrepaid credits, no refunds
GPUBank
Payments and sign-up
Payment methods
A prepaid USD balance, topped up in crypto. Every GPU-second is drawn from it.
Minimum top-up $25
To start
- 1Email and password
- 2Top up in crypto
- 3Launch a GPU
Billing modelPrepaid balance
06Trust and support
Certificates, uptime and who answers.
| GPUBank | ||
|---|---|---|
| SOC 2 | Not published | Not published |
| ISO 27001 | Not published | Not published |
| HIPAA | Not published | No |
| GDPR DPA | Not published | Published DPA, applies automatically |
| Availability target | None, reasonable efforts only | 99.9 % monthly, per region |
| Status page | None published | Public status page |
| Support | Email; sales for 25+ GPUs | Tickets from a $25 balance |
07Their site, our site
What you see before you pay.
Both sites as a visitor sees them, without an account: the JarvisLabs pricing page captured on 29 September 2026, and GPUBank's live board. Cookie banners and chat bubbles are closed; nothing else is edited.

- 1Per-GPU prices are public without login, with a USD and INR toggle and six tabs: on-demand, spot and four reserved terms.
- 2A site-wide banner announces new H200, H100 and RTX PRO 6000 rates from 5 October 2026.
- 3Official pages disagree on the minimum top-up, $5 in the docs and $10 on the home and VM pages.

- 1Live stock per region and every price, without an account.
- 2Per-hour and per-second prices, with vCPU, RAM, NVMe, IPv4 and traffic in them.
- 3A machine-readable price feed at /prices.json.
08The spec sheet
28 points, read off both sites.
GPUBank aheadJarvisLabs aheadEven
| Criterion | GPUBank | |
|---|---|---|
| Price and catalogue | ||
| H100 SXM, per GPU-hour | $1.85, all-in | $2.69, $3.49 from 5 Oct 2026 |
| Cheapest GPU-hour | $0.23, RTX 4090 | $0.41, A30 |
| GPU models listedB200 and B300 are sold only as reserved clusters on quote. | 10 | 9 |
| What the GPU price includes | vCPU, RAM, NVMe, IPv4, traffic | vCPU and RAM; storage billed separately |
| Billing | ||
| Billing granularity | Per second | Per minute |
| Minimum charge | None | None, no minimum rental period |
| Stopped instanceA negative balance deletes instances and data at once, with no grace period. | Disk only | Paused instance pays storage only |
| Persistent storageBilled as 0.00014 USD per GB-hour. Object storage is $0.025 per GB-month. | $0.03 per GB-month | $0.10 per GB-month |
| Egress traffic | Included | Free in-region |
| Public IPv4The included public IP can change on resume. At most 2 static IPs per account. | Included | Public IP included; static IP $5 per month |
| Billing model | Prepaid balance | Prepaid credits, no refunds |
| Payment and sign-up | ||
| Crypto paymentsThe docs FAQ lists credit and debit cards via Stripe only. | Bitcoin, USDT, Ethereum and 12+ more | Not offered |
| Card payments | No | Credit and debit cards via StripeAhead |
| Minimum deposit | $25 | $5 in the docs, $10 on the home pageAhead |
| Platform | ||
| API and CLI | REST API, CLI, Python SDK | CLI and Python SDK |
| Terraform | Official provider | Not published |
| Serverless GPUs | No | Serverless, Managed Endpoints, Model APIsAhead |
| Managed Kubernetes | No | No managed Kubernetes |
| Ready-made templates | 25 | 9 templates |
| Interruptible or spotGPUBank's interruptible H100 SXM costs $1.11. | Yes, 40 % off | Spot, H100 SXM at $1.19 |
| GPUs per instance | Up to 8 | Up to 8, clusters of 128 to 1,024+Ahead |
| Multi-node interconnect | Not offered | NVLink; InfiniBand on clustersAhead |
| Network and regions | ||
| Regions | 4 | 2 (IN1, IN2) |
| Countries | France, Germany, US, Singapore | India |
| EU regions | Paris, Frankfurt | None; EU1 closed to new instances |
| Trust and support | ||
| Availability target | 99.9 % monthly | None, reasonable efforts only |
| Compliance reports | None published | None published |
| Public status page | Yes | None published |
| Support | Tickets from a $25 balance | Email; sales for 25+ GPUs |
09The right fit
Different needs. A clear choice.
What JarvisLabs does better
- Serverless inference, Managed Endpoints and token-priced Model APIs
- Self-serve H200 InfiniBand clusters, reserved clusters of 128 to 1,024+ GPUs
- A30 at $0.41 and L4 at $0.44 per GPU-hour, models GPUBank does not list
- Card payments via Stripe, top-ups from $5 per the docs
- INR billing with GST invoices for Indian companies
Choose GPUBank if
- H100 SXM at $1.85 per GPU-hour, against $2.69 now and $3.49 from 5 Oct
- Per-second billing, with local NVMe, IPv4 and all traffic in the GPU rate
- Regions in Paris, Frankfurt, Ashburn and Singapore, not India only
- Pay in Bitcoin, USDT and 12+ other coins, no card or phone number
- Reserved discount of 20 %, against 2 to 9 % at JarvisLabs
Moving from JarvisLabs takes an afternoon.
JarvisLabs instance disks are tied to their region, and in-region egress and object storage transfer are free. Its VMs run standard Ubuntu, so Docker images and stacks move as they are.
- 1Launch the same GPU on GPUBank with a persistent disk in the region you need.
- 2Copy data with rsync or scp over SSH, or through JarvisLabs object storage with rclone.
- 3Run the same Docker image or pick the matching template, then delete JarvisLabs instances.
# 1. a GPU and a disk on GPUBank gpubank launch --gpu h100-sxm --region eu-paris --disk data # 2. copy your data from the old machine rsync -azP old-host:/workspace/ root@<new-ip>:/workspace/ # 3. run the same image or template ssh root@<new-ip> 'cd /workspace && ./start.sh'
A few things you might be wondering.
Yes. JarvisLabs lists the on-demand H100 SXM at $2.69 per GPU-hour, rising to $3.49 on 5 October 2026, excluding tax and storage. GPUBank charges $1.85 per GPU-hour with vCPU, RAM, NVMe, IPv4 and traffic included. A 40-hour H100 fine-tune costs $107.60 at the current JarvisLabs rate and $74.00 on GPUBank.
Yes. From 5 October 2026 its on-demand H100 SXM goes from $2.69 to $3.49, H200 from $3.99 to $4.59 and RTX PRO 6000 from $1.89 to $2.19. Reserved plans started before that date keep their rate. GPUBank lists the H100 SXM at $1.85, H200 at $2.49 and RTX PRO 6000 at $1.15.
No. Its docs FAQ lists credit and debit cards via Stripe only, and balances are not refundable. GPUBank is paid from a USD balance topped up in Bitcoin, USDT, Ethereum and more than a dozen other coins.
In India only, in regions IN1 and IN2; IN2 is in Noida. Its EU1 region no longer accepts new instances. GPUBank runs in Paris, Frankfurt, Ashburn and Singapore.
JarvisLabs. It offers serverless workers that scale to zero, Managed Endpoints, token-priced Model APIs, self-serve H200 InfiniBand clusters and reserved clusters of 128 to 1,024+ GPUs on quote. GPUBank has no serverless product and rents up to 8 GPUs per instance.
Without. Every price on this page excludes VAT and sales tax, for JarvisLabs and for GPUBank, so the two columns compare like for like. Tax is added at checkout according to where the buyer is based.
Sources
Every figure about JarvisLabs was read on its own pages or official documentation. Prices last checked on 29 September 2026.
- 01jarvislabs.ai/pricingGPU prices all modes, storage, egress, billing
- 02jarvislabs.ai/FAQ: billing, payments, refund, regions, free trial, minimum 10 USD
- 03jarvislabs.ai/updates/gpu-pricing-october-2026Announced price increase
- 04docs.jarvislabs.ai/faqs/Payment methods, storage, balance rules, server location, minimum 5 USD
- 05jarvislabs.ai/termsandserviceBilling, currency, GST, negative balance, refunds, governing law
- 06jarvislabs.ai/supported-countries62 sign-up countries
- 07jarvislabs.ai/signupSign-up fields
- 08jarvislabs.ai/contactusOffices, New Delhi
- 09jarvislabs.ai/products/vmVM configurations, launch times
- 10jarvislabs.ai/products/templatesTemplate catalogue, GPU counts
- 11jarvislabs.ai/products/clustersInstant and reserved clusters, B200/B300
- 12jarvislabs.ai/products/serverlessServerless GPUs, billing
- 13jarvislabs.ai/products/filesystemsNetwork storage price and limits
- 14jarvislabs.ai/products/cliCLI and SDK
- 15jarvislabs.ai/gpu/nvidia-b200B200 by quote, 180 GB
- 16jarvislabs.ai/gpu/nvidia-h200H200 on-demand, spot, reserved
- 17jarvislabs.ai/gpu/nvidia-a100-80gbA100 80GB memory type
- 18jarvislabs.ai/inINR prices, GST invoices, discontinued GPUs
What we could not verify (8)
- Legal entity name and founding year not found on the official pages read.
- A100 80GB form factor (SXM or PCIe) not published; recorded as model 'other'.
- City of region IN1 not published (IN2 is Noida).
- Serverless and Managed Endpoints GPU worker rates not on public pages ('check current GPU rates in the dashboard').
- Reserved plan rates for plans started on or after 5 Oct 2026 not published (by request).
- B200, B300 and reserved-cluster H100/H200 prices not published.
- Terraform provider, snapshots, SLA, status page, compliance certifications not published.
- Whether prices differ by region: GPU pages say regional prices 'vary'; the pricing page shows one rate per GPU.
In short
GPUBank vs JarvisLabs, in facts.
The essentials of this comparison in one table, for quick readers and AI assistants. Prices last checked on 29 September 2026.
- H100 SXM, per GPU-hour
- JarvisLabs: $2.69, $3.49 from 5 Oct 2026
GPUBank: $1.85, all-in - Cheapest GPU-hour
- JarvisLabs: $0.41, A30
GPUBank: $0.23, RTX 4090 - Billing granularity
- JarvisLabs: Per minute
GPUBank: Per second - Stopped instance
- JarvisLabs: Paused instance pays storage only
GPUBank: Disk only - Persistent storage
- JarvisLabs: $0.10 per GB-month
GPUBank: $0.03 per GB-month - Egress traffic
- JarvisLabs: Free in-region
GPUBank: Included - Crypto payments
- JarvisLabs: Not offered
GPUBank: Bitcoin, USDT, Ethereum and 12+ more - Card payments
- JarvisLabs: Credit and debit cards via Stripe
GPUBank: No - Minimum deposit
- JarvisLabs: $5 in the docs, $10 on the home page
GPUBank: $25 - GPU models listed
- JarvisLabs: 9
GPUBank: 10 - Regions
- JarvisLabs: 2 (IN1, IN2)
GPUBank: 4 - Serverless GPUs
- JarvisLabs: Serverless, Managed Endpoints, Model APIs
GPUBank: No - Availability target
- JarvisLabs: None, reasonable efforts only
GPUBank: 99.9 % monthly - Compliance reports
- JarvisLabs: None published
GPUBank: None published - Support
- JarvisLabs: Email; sales for 25+ GPUs
GPUBank: Tickets from a $25 balance - Headquarters
- JarvisLabs: New Delhi, India
GPUBank: Tallinn, Estonia
How this page was made
Every price and policy of JarvisLabs was read on its own website, documentation or public API on 29 September 2026, with the address of each page listed above. A value we could not read on a primary source is marked "Not published" rather than estimated. GPUBank figures come from this site and its terms. Prices change: check both providers before you buy. JarvisLabs, its logo and its website belong to their owner and are shown only to identify the service compared. GPUBank is not affiliated with JarvisLabs.