GPUBank vs GMI Cloud
GMI Cloud is a US GPU cloud, run by GMI Cloud US Inc., that sells dedicated H100, H200, B200 and GB200 NVL72 capacity as containers, bare metal and managed Kubernetes, plus a serverless inference API. Its pricing page shows only 'from' prices: H100 from $2.00 per GPU-hour, form factor not stated; GPUBank lists the H100 SXM at $1.85 all-in. GMI's B200 starts lower and it holds SOC 2 Type 2 and ISO 27001; GPUBank publishes a full price list, bills by the second and launches without sales approval.
GPUBank
Tallinn, Estonia
H200 SXM
$2.49 / h
- vCPU, RAM, NVMe, IPv4 and traffic included
- Per-second billing after the health check
- Top up in Bitcoin, USDT and 12+ other coins
vCPU, RAM, NVMe, IPv4 and traffic included
GMI Cloud
Mountain View, United States
H200 SXM
$2.60 / h
- B200 from $4.00 per GPU-hour, against $4.15 at GPUBank
- GB200 NVL72 available now and GB300 NVL72 on pre-order
- SOC 2 Type 2 report and ISO 27001 certificate in a public trust center
GMI Cloud
GPU cloud
United States22 points compared24 of their pages readprices last checked on 29 September 2026page updated 2 November 2026
The verdict
GB200 racks, compliance and a lower B200 floor at GMI Cloud. Exact all-in prices and self-serve GPUs at GPUBank.
GMI Cloud suits teams that buy H200, B200 or GB200 NVL72 capacity through sales and need SOC 2 Type 2, ISO 27001 and a serverless model API on the same account. Its B200 starts at $4.00 per GPU-hour, below GPUBank's $4.15, while its H200 starts at $2.60 against $2.49. GMI shows only 'from' prices, does not say whether they are on-demand or committed, and bills storage, networking and Elastic IPs separately at rates visible only in the console. GPUBank lists exact all-in prices, bills by the second, and lets any account launch a GPU without a support request.
GMI CloudGPUBank 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 | GMI Cloud per hour | GPUBank per hour | Difference | 100 h at GMI Cloud | 100 h at GPUBank |
|---|---|---|---|---|---|
 B200 SXM180 GB |
$4.00 | $4.15 | +4 % | $400.00 | $415.00 |
 H200 SXM141 GB |
$2.60 | $2.49 | −4 % | $260.00 | $249.00 |
A long-context job, 40 h on an H200
GMI Cloud
$104.00
GPUBank
$99.60
Saves $4.40 at GPUBank (−4 %)
A full day, 24 h on a B200
GMI Cloud
$96.00
GPUBank
$99.60
Costs $3.60 more at GPUBank (+4 %)
02Their catalogue
Every GPU GMI Cloud lists.
5 GPU models listed by GMI Cloud, each at its lowest public price per GPU-hour. 2 of them are also rented at GPUBank (in blue).
GB300reservedNo price- GB200on demand$8.00
- B200on demand$4.00
- H200on demand$2.60
- H100on demand$2.00
03What the price includes
What appears on the bill besides the GPU.
The same H200 SXM session of 2 h 14 m 03 s, line by line, as each provider describes its billing.
GMI Cloud
2 h 14 m 03 s on H200 SXM
- GPU · $2.60 / h
- $5.81
- vCPU and RAM
- Not published
- Storage
- Not published
- Egress
- Not published
- Public IPv4
- Elastic IP extra, price in the console only
- Billing granularity
- Per minute
- Minimum charge
- Not published
GPUBank
2 h 14 m 03 s on H200 SXM
- GPU · $2.49 / h
- $5.56
- 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.
GMI Cloud
Not published.
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 GMI Cloud 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.
United States
Paris
Frankfurt
Ashburn
Singapore
GMI Cloud1
- United States
GPUBank4
Paris
Frankfurt- Ashburn
Singapore
GMI Cloud · 1 countryGPUBank · 4 regions
05Paying and signing up
How each one takes your money.
GMI Cloud
Payments and sign-up
Payment methods
Cards
Bank transfer
No cryptoMinimum top-up Not published
To start
- 1Email and password
- 2Export-control certification
Billing modelPay as you go or prepaid
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.
| GMI Cloud | GPUBank | |
|---|---|---|
| SOC 2 | Type II | Not published |
| ISO 27001 | Listed, report not public | Not published |
| HIPAA | Not published | No |
| GDPR DPA | Not published | Published DPA, applies automatically |
| Availability target | 99.9 % multi-region, 99 % single region | 99.9 % monthly, per region |
| Status page | None found | Public status page |
| Support | Email, Discord, console contact | 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 GMI Cloud pricing page captured on 29 September 2026, and GPUBank's live board. Cookie banners and chat bubbles are closed; nothing else is edited.
- 1The pricing page shows five GPU cards with 'from $X /GPU-hour' and no vCPU, RAM or disk details.
- 2The pricing FAQ says networking and storage are provisioned separately and points to sales for details.
- 3The main calls to action are Contact Sales and Start in Console.
- 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
22 points, read off both sites.
GPUBank aheadGMI Cloud aheadEven
| Criterion | GPUBank | GMI Cloud |
|---|---|---|
| Price and catalogue | ||
| H100 SXM, per GPU-hourGMI does not say if its 'from' prices are on-demand or committed rates. | $1.85, all-in | From $2.00, form factor not stated |
| Cheapest GPU-hour | $0.23, RTX 4090 | From $2.00, H100 |
| GPU models listed | 10 | 5 |
| What the GPU price includesvCPU, RAM and disk per GPU are not stated. | vCPU, RAM, NVMe, IPv4, traffic | Compute only; storage and networking extra |
| Billing | ||
| Billing granularity | Per second | Per minute |
| Minimum chargePrepaid containers cannot be terminated before their rental period ends. | None | Not published |
| Stopped instance | Disk only | Not published |
| Persistent storageShown in the console only. | $0.03 per GB-month | Not published |
| Egress traffic | Included | Not published |
| Public IPv4A disassociated Elastic IP keeps billing until released. | Included | Elastic IP extra, price in the console only |
| Billing model | Prepaid balance | Pay as you go or prepaid |
| Payment and sign-up | ||
| Crypto payments | Bitcoin, USDT, Ethereum and 12+ more | Not published |
| Card payments | No | Saved cards, invoices for contractsAhead |
| Minimum deposit | $25 | Not published |
| Platform | ||
| API and CLI | REST API, CLI, Python SDK | REST API; no CLI found |
| Terraform | Official provider | No provider found |
| Serverless GPUs | No | Inference API, per token or requestAhead |
| Managed Kubernetes | No | Managed GPU clustersAhead |
| Ready-made templates | 25 | 4 built-in plus custom |
| Interruptible or spot | Yes, 40 % off | No |
| GPUs per instanceGB200 NVL72 racks and H200 or B200 clusters are sold through sales. | Up to 8 | Not published |
| Multi-node interconnect | Not offered | RDMA-ready, no bandwidth figureAhead |
| Network and regions | ||
| RegionsMarketing pages mention the US, APAC and EU without naming locations. | 4 | 1 named, us-denver-1 |
| Countries | France, Germany, US, Singapore | United States |
| EU regions | Paris, Frankfurt | None named |
| Trust and support | ||
| Availability targetGMI publishes service credits in GPU-hours. GPUBank's target is 99.9 % per region. | 99.9 % monthly | 99.9 % multi-region, 99 % single region |
| Compliance reports | None published | SOC 2 Type 2, ISO 27001Ahead |
| Public status page | Yes | None found |
| Support | Tickets from a $25 balance | Email, Discord, console contactAhead |
09The right fit
Different needs. A clear choice.
What GMI Cloud does better
- B200 from $4.00 per GPU-hour, against $4.15 at GPUBank
- GB200 NVL72 available now and GB300 NVL72 on pre-order
- SOC 2 Type 2 report and ISO 27001 certificate in a public trust center
- Serverless inference API with 200+ models and dedicated endpoints
- Managed Kubernetes GPU clusters and bare metal with root access
Choose GPUBank if
- H100 SXM at $1.85 all-in, against a 'from $2.00' H100 with extras on top
- A full public price list, not 'from' prices and a console-only rate card
- Launch any GPU yourself; GMI enables bare metal and containers on request
- Per-second billing and an interruptible tier at -40 %
- 10 GPU models including RTX 4090, RTX 5090, L40S and MI300X
Moving from GMI Cloud takes an afternoon.
GMI containers run from templates or custom images, so workloads move as Docker images. Reconfiguring a GMI container wipes its data, so copy data out before any change or termination.
- 1Copy data and models out of the GMI container over SSH or Jupyter to a GPUBank persistent disk.
- 2Run the same Docker image on GPUBank, or pick the vLLM or JupyterLab template.
- 3Terminate GMI containers when the rental period allows, and release Elastic IPs by hand.
# 1. a GPU and a disk on GPUBank gpubank launch --gpu h200-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.
On published prices, yes. GMI lists the H100 from $2.00 per GPU-hour without a form factor, and bills storage and networking separately. GPUBank charges $1.85 per GPU-hour for an H100 SXM and $1.35 for an H100 PCIe, with vCPU, RAM, NVMe, IPv4 and traffic included.
On the 'from' price, yes. GMI lists the B200 from $4.00 per GPU-hour, marked Limited Availability, and GPUBank charges $4.15. A 24-hour day costs $96 at GMI's floor against $99.60 on GPUBank, before GMI's storage and networking.
No. GMI's website shows only 'from' prices, and its docs say exact rates for GPUs, storage and networking appear in the console. GPUBank publishes a full per-GPU-hour price list, all-in, on its website.
Not always. GMI enables bare metal and containers per organization on request to support, and cluster requests are activated by hand. On GPUBank any account with a balance can launch a GPU, and the meter starts after the health check.
GMI does not publish crypto payments. Its docs describe saved cards with auto-pay and invoices for contract customers. GPUBank is paid from a USD balance topped up in Bitcoin, USDT, Ethereum and more than a dozen other coins, from $25.
Without. Every price on this page excludes VAT and sales tax, for GMI Cloud 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 GMI Cloud was read on its own pages or official documentation. Prices last checked on 29 September 2026.
- 01gmicloud.ai/en/pricingGPU prices, pricing FAQ (per GPU-hour, dedicated, networking/storage separate, commitments)
- 02gmicloud.ai/en/gpusGPU catalogue, GB200/GB300 NVL72 labels, product types
- 03gmicloud.ai/enH100/H200 prices, serverless inference
- 04gmicloud.ai/en/company/about30, 000+ GPUs, 99.99% availability claim, SOC 2 / ISO 27001 claim, 24/7 support
- 05gmicloud.ai/en/developersInference API, SDKs, 99.9% uptime target
- 06docs.gmicloud.ai/cluster-engine/indexProduct formats, request flow, product entitlements
- 07docs.gmicloud.ai/cluster-engine/resources/managed-gpu-clustersPer-minute pay-as-you-go billing, prepaid option, manual activation
- 08docs.gmicloud.ai/cluster-engine/resources/containersSSH/Jupyter access, prepaid renewal periods (1 day to 3 years), termination rules
- 09docs.gmicloud.ai/cluster-engine/resources/bare-metalBare metal management, gated access
- 10docs.gmicloud.ai/cluster-engine/resources/elastic-ipIPv4 billed per minute, charges while disassociated
- 11docs.gmicloud.ai/cluster-engine/resources/templatesBuilt-in templates
- 12docs.gmicloud.ai/cluster-engine/user-management/accountAuto-pay with saved card, usage tiers
- 13docs.gmicloud.ai/cluster-engine/user-management/credits-couponsCredit balance, coupons, referral
- 14docs.gmicloud.ai/cluster-engine/user-management/billingOrders and receipts
- 15docs.gmicloud.ai/inference-engine/billing/priceRates shown only in the console
- 16docs.gmicloud.ai/api-reference/introductionREST API, email verification at sign-up, us-denver-1 IDC
- 17docs.gmicloud.ai/inference-engine/marketplace/dedicatedBilling only while Running
- 18gmicloud.ai/en/legal/service-level-agreementSLA figures and service credits
What we could not verify (9)
- H100 form factor (SXM or PCIe) and VRAM for all GPUs: not stated on the pricing or GPUs pages
- Whether the 'from' prices are on-demand or committed rates: not stated
- Reserved or commitment prices and terms: sales only
- Storage, egress and Elastic IP prices: console only, not published
- Stopped-instance billing for bare metal: not documented
- Payment methods beyond saved cards (auto-pay) and invoices; crypto acceptance: not published
- Full region list: only 'us-denver-1' is named in API docs; marketing mentions US, APAC and EU without locations
- HQ city, founding year and funding: not stated on the pages read
- CLI and Terraform provider: not found in the docs index
In short
GPUBank vs GMI Cloud, in facts.
The essentials of this comparison in one table, for quick readers and AI assistants. Prices last checked on 29 September 2026.
Read as Markdown- H100 SXM, per GPU-hour
- GMI Cloud: From $2.00, form factor not stated
GPUBank: $1.85, all-in - Cheapest GPU-hour
- GMI Cloud: From $2.00, H100
GPUBank: $0.23, RTX 4090 - Billing granularity
- GMI Cloud: Per minute
GPUBank: Per second - Stopped instance
- GMI Cloud: Not published
GPUBank: Disk only - Persistent storage
- GMI Cloud: Not published
GPUBank: $0.03 per GB-month - Egress traffic
- GMI Cloud: Not published
GPUBank: Included - Crypto payments
- GMI Cloud: Not published
GPUBank: Bitcoin, USDT, Ethereum and 12+ more - Card payments
- GMI Cloud: Saved cards, invoices for contracts
GPUBank: No - Minimum deposit
- GMI Cloud: Not published
GPUBank: $25 - GPU models listed
- GMI Cloud: 5
GPUBank: 10 - Regions
- GMI Cloud: 1 named, us-denver-1
GPUBank: 4 - Serverless GPUs
- GMI Cloud: Inference API, per token or request
GPUBank: No - Availability target
- GMI Cloud: 99.9 % multi-region, 99 % single region
GPUBank: 99.9 % monthly - Compliance reports
- GMI Cloud: SOC 2 Type 2, ISO 27001
GPUBank: None published - Support
- GMI Cloud: Email, Discord, console contact
GPUBank: Tickets from a $25 balance - Headquarters
- GMI Cloud: Mountain View, United States
GPUBank: Tallinn, Estonia
How this page was made
Every price and policy of GMI Cloud 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. GMI Cloud, its logo and its website belong to their owner and are shown only to identify the service compared. GPUBank is not affiliated with GMI Cloud.
