GPUBank vs Runpod
Runpod is a GPU cloud from San Francisco that sells per-second GPU Pods on a Secure Cloud and a legacy Community Cloud of third-party hosts, plus serverless endpoints and multi-node clusters. Its H100 SXM Pod costs $3.49 per GPU-hour on Secure Cloud and $2.69 on Community Cloud; GPUBank charges $1.85. Runpod has the larger catalogue, serverless, clusters and SOC 2 and ISO 27001; GPUBank costs less on every shared GPU and keeps a stopped instance at a lower disk rate.
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
vCPU, RAM, NVMe, IPv4 and traffic included
Runpod
San Francisco, United States
H100 SXM
$3.49 / h
- 43 GPU models, including B300 and a Community Cloud tier
- Serverless flex workers that scale to zero
- Instant Clusters of 16 to 64 GPUs at 3,200 Gbps between nodes
Runpod
GPU cloud
United States27 points compared32 of their pages readprices last checked on 29 September 2026page updated 2 November 2026
The verdict
Catalogue, serverless and clusters at Runpod. Lower Pod prices at GPUBank.
Runpod and GPUBank both bill by the second and include traffic. For the nine GPUs both of us rent, GPUBank costs 39 to 69 % less than Runpod's Secure Cloud list price. Runpod goes further on scope: 43 GPU models, serverless workers that scale to zero, instant clusters of 16 to 64 GPUs, 11 countries and SOC 2 Type II with ISO 27001. Runpod takes cards and crypto; GPUBank takes crypto only.
RunpodGPUBank 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 | Runpod per hour | GPUBank per hour | Difference | 100 h at Runpod | 100 h at GPUBank |
|---|---|---|---|---|---|
 B200 SXM180 GB |
$6.79 | $4.15 | −39 % | $679.00 | $415.00 |
 H200 SXM141 GB |
$4.31 | $2.49 | −42 % | $431.00 | $249.00 |
 H100 SXM80 GB |
$3.49 | $1.85 | −47 % | $349.00 | $185.00 |
 H100 PCIe80 GB |
$2.89 | $1.35 | −53 % | $289.00 | $135.00 |
 RTX PRO 600096 GB |
$2.09 | $1.15 | −45 % | $209.00 | $115.00 |
 A100 SXM80 GB |
$1.59 | $0.95 | −40 % | $159.00 | $95.00 |
 L40S48 GB |
$1.09 | $0.55 | −50 % | $109.00 | $55.00 |
 RTX 509032 GB |
$0.99 | $0.48 | −52 % | $99.00 | $48.00 |
 RTX 409024 GB |
$0.74 | $0.23 | −69 % | $74.00 | $23.00 |
A fine-tune, 40 h on an H100 SXM
Runpod
$139.60
GPUBank
$74.00
Saves $65.60 at GPUBank (−47 %)
An inference box, a month, 720 h, on an L40S
Runpod
$784.80
GPUBank
$396.00
Saves $388.80 at GPUBank (−50 %)
Image generation, 200 h on an RTX 4090
Runpod
$148.00
GPUBank
$46.00
Saves $102.00 at GPUBank (−69 %)
02Their catalogue
Every GPU Runpod lists.
Runpod lists 43 GPU models. The 26 below are the ones recorded in our check, each at its lowest public price per GPU-hour. 9 of them are also rented at GPUBank (in blue).
- Blackwell and Hopper6
- B300288 GB · on demand$7.89
- B200180 GB · on demand$6.79
- H200on demand$4.31
- H100 SXM80 GB · on demand$3.49
- H100 NVL94 GB · on demand$3.19
- H100 PCIe80 GB · on demand$2.89
- Ampere, Ada and older data-centre6
- A100 SXM 80GB80 GB · on demand$1.59
A100 PCIe 80GB80 GB · on demand$1.59- L40S48 GB · on demand$1.09
- Ampere, Ada and older data-centre continued6
- L4048 GB · on demand$0.82
L424 GB · on demand$0.49
A4048 GB · on demand$0.49- Workstation11
- RTX PRO 600096 GB · on demand$2.09
RTX 6000 Ada48 GB · on demand$0.84
RTX A600048 GB · on demand$0.53
RTX A500024 GB · on demand$0.27
RTX A4000serverless$0.58- RTX PRO 4500 Blackwellserverless$1.15
- Workstation continued11
RTX 2000 Adaserverless$0.58- RTX 4000serverless$0.58
- RTX A4500serverless$0.58
- RTX PRO 6000 MIG 24GB24 GB · on demandNo price
- RTX PRO 6000 MIG 48GB48 GB · on demandNo price
- Consumer3
- RTX 509032 GB · on demand$0.99
- RTX 409024 GB · on demand$0.74
RTX 309024 GB · on demand$0.50
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.
Runpod
2 h 14 m 03 s on H100 SXM
- GPU · $3.49 / h
- $7.80
- vCPU and RAM
- Included
- Storage
- $0.10 volume, $0.07 network per GB-month
- Egress
- Included
- Public IPv4
- Included on Secure Cloud, optional on Community
- Billing granularity
- Per second
- Minimum charge
- None per launch; one hour of credits to deploy
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.
Runpod
Volume disk at $0.20 per GB-month.
Network volumes keep billing. Container disk is erased on stop.
GPUBank
Only the persistent disk, $0.03 per GB-month. Compute, IPv4 and scratch cost nothing.
04Where the GPUs run
11 countries against 4, and where they are.
Countries where Runpod 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.
Paris
Frankfurt
Ashburn
Singapore
Runpod11
Australia
Canada
Czech Republic
France
Iceland
Japan
Netherlands
Norway
Romania
Sweden- United States
GPUBank4
- Paris
Frankfurt- Ashburn
Singapore
Runpod · 11 countriesGPUBank · 4 regions
05Paying and signing up
How each one takes your money.
Runpod
Payments and sign-up
Payment methods
Cards
Bank transfer
CryptoCrypto: Accepted
Minimum top-up $10 per the docs
To start
- 1Email
- 2Prepaid credits before deploying
Billing modelPrepaid credits; post-paid for enterprise
GPUBank
Payments and sign-up
Payment methods
No cardsA 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.
| Runpod | GPUBank | |
|---|---|---|
| SOC 2 | Type II | Not published |
| ISO 27001 | Yes | Not published |
| HIPAA | BAA available | No |
| GDPR DPA | DPA on request | Published DPA, applies automatically |
| Availability target | 99.99% in enterprise agreements | 99.9 % monthly, per region |
| Status page | uptime.runpod.io | Public status page |
| Support | Portal, email, Discord; TAM for enterprise | 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 Runpod 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 hourly Pod prices are public without an account and default to the Secure Cloud tab, with a toggle to Community Cloud.
- 2Serverless shows only flex worker rates; active worker pricing goes through sales.
- 3Cluster prices are listed for H200 SXM and A100 SXM only; other GPUs and all reserved terms say Contact sales.
- 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
27 points, read off both sites.
GPUBank aheadRunpod aheadEven
| Criterion | GPUBank | Runpod |
|---|---|---|
| Price and catalogue | ||
| H100 SXM, per GPU-hour | $1.85, all-in | $3.49 Secure Cloud, $2.69 Community Cloud |
| Cheapest GPU-hourSecure Cloud starts at $0.27 for the same card. Our entry is an RTX 4090 at $0.23. | $0.23, RTX 4090 | $0.16, RTX A5000 on Community CloudAhead |
| GPU models listed | 10 | 43Ahead |
| What the GPU price includesContainer and volume disks cost $0.10 per GB-month while a Pod runs. | vCPU, RAM, NVMe, IPv4, traffic | vCPU and RAM; disks billed apart |
| Billing | ||
| Billing granularity | Per second | Per second |
| Minimum chargeDeploying an on-demand Pod requires at least one hour's worth of credits in the balance. | None | None per launch; one hour of credits to deploy |
| Stopped instanceNetwork volumes keep billing. Container disk is erased on stop. | Disk only | Volume disk at $0.20 per GB-month |
| Persistent storageNetwork volumes drop to $0.05 per GB-month over 1 TB. Volume disk costs $0.20 when stopped. | $0.03 per GB-month | $0.10 volume, $0.07 network per GB-month |
| Egress traffic | Included | Included |
| Public IPv4 | Included | Included on Secure Cloud, optional on Community |
| Billing model | Prepaid balance | Prepaid credits; post-paid for enterpriseAhead |
| Payment and sign-up | ||
| Crypto paymentsCrypto.com is named in the terms. The coin list is not published. | Bitcoin, USDT, Ethereum and 12+ more | Accepted |
| Card payments | No | Cards via StripeAhead |
| Minimum depositPrepaid cards should deposit at least $100 per transaction. | $25 | $10 per the docsAhead |
| Platform | ||
| API and CLI | REST API, CLI, Python SDK | REST and GraphQL API, runpodctl CLI, SDKs |
| Terraform | Official provider | Official provider |
| Serverless GPUs | No | Yes, flex workers that scale to zeroAhead |
| Managed Kubernetes | No | Not published |
| Ready-made templates | 25 | Not published |
| Interruptible or spot | Yes, 40 % off | Interruptible Pods, no price published |
| GPUs per instanceMaximum GPUs per Pod is not published. | Up to 8 | Clusters of 16 to 64 GPUs, 512 via salesAhead |
| Multi-node interconnect1,600 Gbps for A100 clusters. | Not offered | 3,200 Gbps between cluster nodesAhead |
| Network and regions | ||
| Regions | 4 | 28 data centers, 31 claimedAhead |
| Countries | France, Germany, US, Singapore | 11, from the US to Japan and AustraliaAhead |
| EU regions | Paris, Frankfurt | France, Netherlands, Sweden, Romania and moreAhead |
| Trust and support | ||
| Availability targetThe home page advertises 99.9% uptime. Community Cloud has no uptime warranty. | 99.9 % monthly | 99.99% in enterprise agreements |
| Compliance reports | None published | SOC 2 Type II, ISO 27001, HIPAA programAhead |
| Public status page | Yes | uptime.runpod.io |
| Support | Tickets from a $25 balance | Portal, email, Discord; TAM for enterpriseAhead |
09The right fit
Different needs. A clear choice.
What Runpod does better
- 43 GPU models, including B300 and a Community Cloud tier
- Serverless flex workers that scale to zero
- Instant Clusters of 16 to 64 GPUs at 3,200 Gbps between nodes
- SOC 2 Type II, ISO 27001, SOC 3 and a HIPAA program
- Card payments, invoicing and 28 data centers in 11 countries
Choose GPUBank if
- H100 SXM at $1.85 per GPU-hour, against $3.49 on Secure Cloud
- RTX 4090 at $0.23, against $0.74 on Secure Cloud
- Local NVMe included in the GPU price
- Stopped instance disk at $0.03 per GB-month, against $0.20
Moving from Runpod takes an afternoon.
Runpod workloads are standard Docker containers and Runpod charges no egress, so images and data move out as they are. Serverless handlers use the Runpod SDK and need a replacement queue layer.
- 1Launch the same GPU on GPUBank with a persistent disk in the region you need.
- 2Copy volume data with rsync over SSH, runpodctl send, or the S3-compatible API on network volumes.
- 3Run the same Docker image on GPUBank, then stop and delete the Runpod Pods and volumes.
# 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. Runpod lists the H100 SXM Pod at $3.49 per GPU-hour on Secure Cloud and $2.69 on Community Cloud. GPUBank charges $1.85 per GPU-hour, with vCPU, RAM, NVMe, IPv4 and traffic included.
GPU and CPU stop billing, but the volume disk is billed at $0.20 per GB-month, double the running rate, and network volumes keep billing. On GPUBank a stopped instance costs only its persistent disk, at $0.03 per GB-month.
No. Runpod runs serverless endpoints with flex workers that scale to zero and bills them per second. GPUBank rents instances only, so for bursty inference that idles most of the day Runpod Serverless can cost less.
Runpod. Its Instant Clusters run 16 to 64 GPUs on demand with 3,200 Gbps between nodes, and Reserved Clusters go further on contract. GPUBank rents up to 8 GPUs per instance and does not sell clusters.
Sources
Every figure about Runpod was read on its own pages or official documentation. Prices last checked on 29 September 2026.
- 01runpod.io/pricingPod Secure/Community prices, serverless flex prices, cluster prices, storage prices
- 02docs.runpod.io/pods/pricingsavings plans, storage pricing, minimum balance, spend limit
- 03docs.runpod.io/serverless/pricingflex vs active workers, per-second billing
- 04docs.runpod.io/accounts-billing/billingprepaid credits, payment methods, invoicing threshold, refunds, minimum deposit
- 05docs.runpod.io/accounts-billing/post-paid-billingenterprise invoicing
- 06docs.runpod.io/accounts-billing/manage-payment-cardsStripe cards, crypto and ACH/wire alternatives
- 07docs.runpod.io/accounts-billing/manage-accountssign-up, team roles, audit logs
- 08docs.runpod.io/references/gpu-typescatalogue count, GPU pools
- 09docs.runpod.io/pods/choose-a-podSecure vs Community Cloud, MIG availability, Community hosts closed
- 10docs.runpod.io/pods/overviewtemplates, SSH/Jupyter, limitations
- 11docs.runpod.io/pods/storage/typescontainer/volume/network/global volume prices, encryption
- 12docs.runpod.io/storage/network-volumesnetwork volume price, Secure Cloud only, migration
- 13docs.runpod.io/storage/high-performance-storagepremium tier pricing policy
- 14docs.runpod.io/storage/globalvolume/overviewrequest charges, deletion after 15 days at zero balance
- 15docs.runpod.io/instant-clusterscluster sizes and interconnect
- 16docs.runpod.io/api-reference/pods/POST/podsinterruptible Pods, public IP
- 17docs.runpod.io/api-reference/overviewAPI v1 deprecation, OpenAPI
- 18docs.runpod.io/runpodctl/overviewCLI capabilities
What we could not verify (14)
- Savings plan (3- or 6-month prepaid) prices: only shown in the console
- Interruptible/spot Pod prices: the API supports interruptible Pods but no spot price is published on the pricing page
- Serverless active worker prices: 'discounts available through sales inquiry'
- Cluster prices for L40S, H100 SXM, B200 and all reserved cluster terms: 'Contact sales'
- Instant Cluster rates are shown as '$/hr' without the words 'per GPU'; per-GPU basis inferred (see row notes)
- MI300X, H200 NVL, V100, RTX 30/40-series consumer cards and other GPU-types entries have no public price on the pricing page
- Community Cloud values for the two MIG slices exist in page data (1.00 and 0.50) but docs say MIG is Secure Cloud only; not recorded
- Pods overview doc says Pods are 'billed by the minute' while pricing docs and product page say per second
- High-performance storage: pricing page lists 0.14 USD/GB-month, docs say price varies by data center
- Maximum GPUs per Pod not published in the pages read
- Founding year and funding not published on the pages read
- Free credits: home page says 'Get started for free' but no credit amount is published
- Managed Kubernetes, snapshots, bare metal: not documented in the pages read
- Crypto coins and rails not listed
In short
GPUBank vs Runpod, 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
- Runpod: $3.49 Secure Cloud, $2.69 Community Cloud
GPUBank: $1.85, all-in - Cheapest GPU-hour
- Runpod: $0.16, RTX A5000 on Community Cloud
GPUBank: $0.23, RTX 4090 - Billing granularity
- Runpod: Per second
GPUBank: Per second - Stopped instance
- Runpod: Volume disk at $0.20 per GB-month
GPUBank: Disk only - Persistent storage
- Runpod: $0.10 volume, $0.07 network per GB-month
GPUBank: $0.03 per GB-month - Egress traffic
- Runpod: Included
GPUBank: Included - Crypto payments
- Runpod: Accepted
GPUBank: Bitcoin, USDT, Ethereum and 12+ more - Card payments
- Runpod: Cards via Stripe
GPUBank: No - Minimum deposit
- Runpod: $10 per the docs
GPUBank: $25 - GPU models listed
- Runpod: 43
GPUBank: 10 - Regions
- Runpod: 28 data centers, 31 claimed
GPUBank: 4 - Serverless GPUs
- Runpod: Yes, flex workers that scale to zero
GPUBank: No - Availability target
- Runpod: 99.99% in enterprise agreements
GPUBank: 99.9 % monthly - Compliance reports
- Runpod: SOC 2 Type II, ISO 27001, HIPAA program
GPUBank: None published - Support
- Runpod: Portal, email, Discord; TAM for enterprise
GPUBank: Tickets from a $25 balance - Headquarters
- Runpod: San Francisco, United States
GPUBank: Tallinn, Estonia
How this page was made
Every price and policy of Runpod 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. Runpod, its logo and its website belong to their owner and are shown only to identify the service compared. GPUBank is not affiliated with Runpod.
