GPUBank vs Prime Intellect
Prime Intellect is a US AI company whose GPU marketplace resells on-demand instances from other clouds, such as Lambda Cloud, Massed Compute, Nebius and RunPod, through one prepaid account. It listed no single-node H100 SXM at capture; its H100 PCIe offers read $2.35 and $3.29 per GPU-hour on two loads, against $1.35 for an H100 PCIe and $1.85 for an H100 SXM at GPUBank. Prime Intellect adds multi-node InfiniBand clusters and card payments; GPUBank publishes fixed all-in prices, bills by the second and keeps the disk of a stopped instance.
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
Prime Intellect
United States
H200 SXM
$5.33 / h
- Multi-node H100 SXM clusters of up to 256 GPUs, 3.2 Tbps InfiniBand on Premium
- 13 GPU models across partner clouds, including A10, L40 and RTX 6000 Ada
- Card payments through Stripe, plus crypto through Crypto.com
Prime Intellect
Aggregator
24 points compared8 of their pages readprices last checked on 29 September 2026page updated 2 November 2026
The verdict
Clusters and a wider catalogue at Prime Intellect. Fixed, lower prices and per-second billing at GPUBank.
Prime Intellect suits teams that want one account across several GPU clouds, H100 clusters of up to 256 GPUs and its RL training stack. Its prices are set by each partner cloud and changed between two page loads during our capture. At that time, GPUBank cost 23 to 53 % less for the GPUs both of us rent, with vCPU, RAM, disk, IPv4 and traffic in the price. Both accept crypto: Prime Intellect through Crypto.com, with the coins not listed, and it also takes cards through Stripe.
Prime IntellectGPUBank 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 | Prime Intellect per hour | GPUBank per hour | Difference | 100 h at Prime Intellect | 100 h at GPUBank |
|---|---|---|---|---|---|
 H200 SXM141 GB |
$5.33 | $2.49 | −53 % | $533.00 | $249.00 |
 H100 PCIe80 GB |
$2.35 | $1.35 | −43 % | $235.00 | $135.00 |
 A100 SXM80 GB |
$1.23 | $0.95 | −23 % | $123.00 | $95.00 |
 L40S48 GB |
$0.82 | $0.55 | −33 % | $82.00 | $55.00 |
Prime Intellect prices are a live marketplace snapshot read on 2026-09-29, set by partner clouds, and they changed between page loads. The H200 figure is GPU only; CPU and RAM are billed on top.
An inference box, a month, 720 h, on an L40S
Prime Intellect
$590.40
GPUBank
$396.00
Saves $194.40 at GPUBank (−33 %)
A long-context job, 40 h on an H200
Prime Intellect
$213.20
GPUBank
$99.60
Saves $113.60 at GPUBank (−53 %)
A training run, 100 h on an A100 SXM
Prime Intellect
$123.00
GPUBank
$95.00
Saves $28.00 at GPUBank (−23 %)
02Their catalogue
Every GPU Prime Intellect lists.
Prime Intellect lists 13 GPU models. The 11 below are the ones recorded in our check, each at its lowest public price per GPU-hour. 5 of them are also rented at GPUBank (in blue).
- H200141 GB · on demand$5.33
H100 SXM80 GB · on demandNo price- H100 PCIe80 GB · on demand$2.35
- A100 SXM 80GB80 GB · on demand$1.23
A100 PCIe 80GB80 GB · on demand$1.20- A100 40GB40 GB · on demand$1.99
- L40S48 GB · on demand$0.82
- L4048 GB · on demand$0.86
A1024 GB · on demand$1.29
RTX 6000 Ada48 GB · on demand$0.75
RTX A600048 GB · on demand$0.54
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.
Prime Intellect
2 h 14 m 03 s on H200 SXM
- GPU · $5.33 / h
- $11.91
- vCPU and RAM
- Not published
- Storage
- $0.00014 per GB-hour on RunPod offers
- Egress
- Not published
- Public IPv4
- Not published
- Billing granularity
- Per minute, from prepaid credits
- 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.
Prime Intellect
Pause on RunPod offers only; data lost on termination.
GPUBank
Only the persistent disk, $0.03 per GB-month. Compute, IPv4 and scratch cost nothing.
04Where the GPUs run
3 countries against 4, and where they are.
Countries where Prime Intellect 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
Canada
Finland
Paris
Frankfurt
Ashburn
Singapore
Prime Intellect3
United States
Canada
Finland
GPUBank4
Paris
Frankfurt- Ashburn
Singapore
Prime Intellect · 3 countriesGPUBank · 4 regions
05Paying and signing up
How each one takes your money.
Prime Intellect
Payments and sign-up
Payment methods
Cards
CryptoCrypto: Yes, via Crypto.com; coins not listed
Minimum top-up Not published
To start
- 1Account (login required to deploy)
- 2Payment via Stripe or Crypto.com
Billing modelPrepaid credits
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.
| Prime Intellect | 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 | No formal SLA | 99.9 % monthly, per region |
| Status page | Linked in the site footer | Public status page |
| Support | Chat, Discord, email | 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 Prime Intellect home page captured on 4 October 2026, and GPUBank's live board. Cookie banners and chat bubbles are closed; nothing else is edited.
- 1www.primeintellect.ai/pricing returns 404; GPU prices sit on the app's on-demand page, visible without login.
- 2Prices are labeled as updating in real time, and each offer names the underlying provider.
- 3A spot toggle is shown, but no spot offers appeared at capture time.
- 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
24 points, read off both sites.
GPUBank aheadPrime Intellect aheadEven
| Criterion | GPUBank | Prime Intellect |
|---|---|---|
| Price and catalogue | ||
| H100 SXM, per GPU-hourLive offers. GPUBank: H100 PCIe $1.35, H100 SXM $1.85. | $1.85, all-in | No H100 SXM; H100 PCIe $2.35 to $3.29 |
| Cheapest GPU-hour | $0.23, RTX 4090 | $0.54, RTX A6000 |
| GPU models listed | 10 | 13Ahead |
| What the GPU price includes | vCPU, RAM, NVMe, IPv4, traffic | Varies by provider; some add CPU, RAM, storage |
| Billing | ||
| Billing granularity | Per second | Per minute, from prepaid credits |
| Minimum charge | None | Not published |
| Stopped instance | Disk only | Pause on RunPod offers only; data lost on termination |
| Persistent storageSet by each provider; other offers not published. | $0.03 per GB-month | $0.00014 per GB-hour on RunPod offers |
| Egress traffic | Included | Not published |
| Public IPv4 | Included | Not published |
| Billing model | Prepaid balance | Prepaid credits |
| Payment and sign-up | ||
| Crypto payments | Bitcoin, USDT, Ethereum and 12+ more | Yes, via Crypto.com; coins not listed |
| Card payments | No | Yes, via StripeAhead |
| Minimum deposit | $25 | Not published |
| Platform | ||
| API and CLI | REST API, CLI, Python SDK | REST API, prime CLI via pip |
| Terraform | Official provider | No provider published |
| Serverless GPUs | No | Serverless inferenceAhead |
| Managed Kubernetes | No | No |
| Ready-made templates | 25 | Not published |
| Interruptible or spotThe FAQ describes spot discounts of up to 90 %. | Yes, 40 % off | Toggle shown, no offers at capture |
| GPUs per instance | Up to 8 | Up to 8; clusters up to 256 GPUsAhead |
| Multi-node interconnect | Not offered | 3.2 Tbps InfiniBand on Premium clustersAhead |
| Network and regions | ||
| Regions | 4 | Offers in 3 countries |
| Countries | France, Germany, US, Singapore | US, Canada, Finland |
| EU regions | Paris, Frankfurt | Finland, on Nebius offers |
| Trust and support | ||
| Availability target | 99.9 % monthly | No formal SLA |
| Compliance reports | None published | None published |
| Public status page | Yes | Linked in the site footer |
| Support | Tickets from a $25 balance | Chat, Discord, emailAhead |
09The right fit
Different needs. A clear choice.
What Prime Intellect does better
- Multi-node H100 SXM clusters of up to 256 GPUs, 3.2 Tbps InfiniBand on Premium
- 13 GPU models across partner clouds, including A10, L40 and RTX 6000 Ada
- Card payments through Stripe, plus crypto through Crypto.com
- Serverless inference, code sandboxes and hosted RL training
- RTX A6000 offers from $0.54 per GPU-hour
Choose GPUBank if
- H200 at $2.49 per GPU-hour, against $5.33 plus CPU and RAM
- Fixed published prices, not a live marketplace snapshot
- Per-second billing instead of per-minute credit deduction
- Stop an instance and keep its disk; Prime pauses only on RunPod offers
- A 99.9 % monthly SLA target; Prime Intellect states no formal SLA
Moving from Prime Intellect takes an afternoon.
Prime Intellect instances run on partner clouds, and terminated instances cannot be recovered. Pods are also deleted when credits run out, so copy data out first.
- 1Launch the same GPU on GPUBank with a persistent disk in the region you need.
- 2Copy checkpoints and datasets from the Prime Intellect instance with rsync over SSH, or to S3.
- 3Run the same Docker image or pick a matching template, then terminate the Prime Intellect instance.
# 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.
Yes, at the time of our capture. Prime Intellect listed no single-node H100 SXM. Its H100 PCIe offers read $2.35 from Massed Compute and $3.29 from Lambda Cloud on two loads of the same page. GPUBank charges a fixed $1.35 for an H100 PCIe and $1.85 for an H100 SXM.
Yes. Its terms list payment through Stripe and Crypto.com, but the accepted coins and the minimum deposit are not published. GPUBank is paid from a USD balance topped up in Bitcoin, USDT, Ethereum and more than a dozen other coins, from $25, and takes no cards.
Terminated instances lose all data, and pause and resume exist only on RunPod offers. When credits run out, pods are deleted automatically. On GPUBank a stopped instance keeps its persistent disk, billed at $0.03 per GB-month.
Prime Intellect. It spins up H100 SXM5 clusters of up to 256 GPUs on demand, with 3.2 Tbps InfiniBand on the Premium configuration and Slurm guides. Its docs quote about $52.80 per hour for Premium and $40.80 for Value, without the GPUs per node. GPUBank rents up to 8 GPUs per instance and sells no InfiniBand clusters.
It is a marketplace: each partner cloud sets the hourly price of its offers, and the page says prices update in real time. Some offers add charges, for example the Nebius H200 adds $0.012 per CPU core-hour and $0.0032 per GB-hour of RAM. GPUBank prices are fixed and include vCPU, RAM, NVMe, IPv4 and traffic.
Without. Every price on this page excludes VAT and sales tax, for Prime Intellect 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 Prime Intellect was read on its own pages or official documentation. Prices last checked on 29 September 2026.
- 01app.primeintellect.ai/dashboard/on-demand-gpusGPU cards with cheapest live prices and availability
- 02app.primeintellect.ai/dashboard/on-demand-gpus/offersPer-provider live offers, socket, location
- 03docs.primeintellect.ai/faq.mdPer-minute credit deduction, spot definition, no SLA, pause, multi-node cluster prices, refunds
- 04primeintellect.ai/terms-of-servicePayment via Stripe and Crypto.com, no refunds, Delaware law
- 05docs.primeintellect.ai/api-reference/check-gpu-availability.mdAvailability/pricing API
- 06primeintellect.ai/securityVulnerability disclosure, no certifications listed
- 07primeintellect.ai/Products, investors, CLI
- 08primeintellect.ai/pricingConfirms no marketing pricing page
What we could not verify (7)
- Prices are a real-time snapshot; they changed between two loads (H100 PCIe 2.35 vs 3.29). No stable price list published.
- Spot offer prices not visible at capture (FAQ says spot discounts up to 90%).
- Multi-node cluster prices '~$52.80/hr' and '~$40.80/hr' from the docs FAQ lack GPUs-per-node; no per-GPU value computed.
- Offers for other GPU types (VRAM filters list 16, 20, 24, 32, 96 GB classes) not shown as available; full provider catalogue not published.
- Crypto coins accepted via Crypto.com not listed; minimum deposit not published.
- Company HQ city and founding year not stated on pages read.
- Egress and IPv4 policy depend on the underlying provider; not published.
In short
GPUBank vs Prime Intellect, 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
- Prime Intellect: No H100 SXM; H100 PCIe $2.35 to $3.29
GPUBank: $1.85, all-in - Cheapest GPU-hour
- Prime Intellect: $0.54, RTX A6000
GPUBank: $0.23, RTX 4090 - Billing granularity
- Prime Intellect: Per minute, from prepaid credits
GPUBank: Per second - Stopped instance
- Prime Intellect: Pause on RunPod offers only; data lost on termination
GPUBank: Disk only - Persistent storage
- Prime Intellect: $0.00014 per GB-hour on RunPod offers
GPUBank: $0.03 per GB-month - Egress traffic
- Prime Intellect: Not published
GPUBank: Included - Crypto payments
- Prime Intellect: Yes, via Crypto.com; coins not listed
GPUBank: Bitcoin, USDT, Ethereum and 12+ more - Card payments
- Prime Intellect: Yes, via Stripe
GPUBank: No - Minimum deposit
- Prime Intellect: Not published
GPUBank: $25 - GPU models listed
- Prime Intellect: 13
GPUBank: 10 - Regions
- Prime Intellect: Offers in 3 countries
GPUBank: 4 - Serverless GPUs
- Prime Intellect: Serverless inference
GPUBank: No - Availability target
- Prime Intellect: No formal SLA
GPUBank: 99.9 % monthly - Compliance reports
- Prime Intellect: None published
GPUBank: None published - Support
- Prime Intellect: Chat, Discord, email
GPUBank: Tickets from a $25 balance - Headquarters
- Prime Intellect: United States
GPUBank: Tallinn, Estonia
How this page was made
Every price and policy of Prime Intellect 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. Prime Intellect, its logo and its website belong to their owner and are shown only to identify the service compared. GPUBank is not affiliated with Prime Intellect.
