Pricing
Every price, before you start.
AI models per million tokens, and GPUs per hour with each class's billing minimum beside the price, all before VAT. Self-serve runs on a prepaid balance; universities and contracts pay by invoice.
From €0,1575 per 1M input tokens · Mistral Small 3.2, EU-only
AI Gateway
AI models by the token.
Input and output per 1M tokens, excl. VAT, grouped by where requests are processed.
Listed prices are the standard rate, with a 5 percent markup. Requests admitted before an account reaches 5.000.000 tokens in a calendar month are charged at cost.
| Model | Price / 1M tokens | Try |
|---|---|---|
|
EU-only
Paris, France · EU-headquartered provider · no fallback outside the EU |
||
| DeepSeek V4 Flash | In €0,42 · Out €0,84 | Try |
| Gemma 4 26B | In €0,2625 · Out €0,525 | Try |
| GLM 5.2 | In €1,89 · Out €5,775 | Try |
| GPT-OSS 120B | In €0,1575 · Out €0,63 | Try |
| Llama 3.3 70B | In €0,945 · Out €0,945 | Try |
| Mistral Medium 3.5 | In €1,575 · Out €7,875 | Try |
| Mistral Small 3.2 | In €0,1575 · Out €0,3675 | Try |
| Qwen3 235B | In €0,7875 · Out €2,3625 | Try |
| Qwen3.5 397B | In €0,63 · Out €3,78 | Try |
| Qwen3.6 35B | In €0,2625 · Out €1,575 | Try |
| Qwen3.8 27B | In €0,63 · Out €3,465 | Try |
| May be processed outside the EU | ||
| Claude Opus 5.5 | In €7,73 · Out €38,64 | Try |
| Gemini 4 Argon | In €3,48 · Out €17,39 | Try |
| Gemini 3.8 Flash | In €1,31 · Out €6,53 | Try |
razorBridge Managed kOS
Shared or Dedicated, with Outpost on site
Priced with you, invoiced by contract
AI Blades
GPUs by the hour.
Prices per GPU per hour, excl. VAT. Each row shows the minimum a session bills.
| GPU | VRAM | Best for | Minimum per session | € per hour, excl. VAT |
|---|---|---|---|---|
| RTX 4000 Ada | 20 GB | Light workloads & demos | Not available in the EU | |
| L4 | 24 GB | Light workloads & demos | 60 min | €1,09 |
| L40S | 48 GB | Training & workshops | 60 min | €1,99 |
| RTX 6000 Ada | 48 GB | Training & workshops | Not available in the EU | |
| MI300X | 192 GB | Large model inference | Not available in the EU | |
| H100 | 80 GB | Large model training | 60 min | €4,29 |
| H200 | 141 GB | Large model training | Not available in the EU | |
| L40S x4Sold out in the EU right now | 192 GB | Large model training | 60 min | €7,99 |
| L40S x8Sold out in the EU right now | 384 GB | Distributed training | 60 min | €15,99 |
| MI300X x8 | 1536 GB | Multi-GPU inference | Not available in the EU | |
| H100 x8Sold out in the EU right now | 640 GB | Multi-GPU workloads | 60 min | €29,99 |
| H200 x8 | 1128 GB | Multi-GPU workloads | Not available in the EU | |
Unified-memory Macs for private inference.
How you pay
Two ways to pay, one price list.
Self-serve accounts draw on a prepaid balance. Universities and organisations under contract pay by invoice.
Self-serve
Account
€0 to open an account
A prepaid balance. Top it up by card in the Hub, or ask us for a pro-forma invoice.
- No subscription or monthly fee
- Every GPU priced above
- AI Gateway access
- SSH and REST API access
Pay as you go
Pay-as-you-go
from €1,09/h
Charged from your credit balance at the hourly rate, excl. VAT. Each session bills at least 60 minutes, then per second.
- 60-minute minimum, then per second
- More credits on request
- Multi-user events
- Usage dashboard
Universities and contracts
By invoice
Pro-forma then an invoice
Pro-forma, then an invoice from KlusAI Labs SRL, with reverse charge for EU VAT-registered institutions.
- A budget agreed before the course starts
- Labs and events run from the Hub
- Students never enter a card
Events and labs
A GPU lab for your class, in one link.
For courses, workshops and hackathons, built on AI Blades. You plan it, students join, and it cleans up after itself.
GPU 1L4
GPU 2L4
GPU 3L4
GPU 4L4
GPU 5L4
GPU 6L4
Illustrative live board
01 Plan
See the total first.GPU, seats and hours. The estimate is the event's cost before VAT.
02 Share
A link and a QR code.Students sign in and join. Nobody enters a card.
03 Done
It tears itself down.GPUs stop at the end time.
What will it cost?
Status
GPU availability
Live availability status across our GPU fleet.
Questions
Straight answers.
Something else? Talk to us.
How does billing work?
AI Blades are charged against your credit balance at the hourly rates above, excl. VAT. AI Gateway is charged per token. Usage is deducted as it happens, and you can follow it in the Hub.
One exception: some EU GPUs bill a 60-minute minimum. This applies to L4, L40S and H100 GPUs, including their multi-GPU versions, when they run in the EU: a session is charged for at least 60 minutes, then per second. The Hub shows it before you start a session.
What happens when credits run out?
We email you when your balance falls below about an hour of runtime. When it runs out, a self-serve session stops automatically, and anything not saved off the blade is lost. To keep going, top up your balance by card in the Hub, or ask us for a pro-forma invoice.
Can I get a refund for unused credits?
Unused credit you paid for can be refunded within 14 days of purchase. Compute you already used cannot be refunded. See the Refund Policy.
What's the minimum spend?
There is no subscription or monthly fee. Each GPU session bills at least 60 minutes, then per second, at the hourly rate excl. VAT. Top up your balance by card in the Hub, or ask us for a pro-forma invoice.
How do I get more credits?
Top up your balance by card in the Hub, or email [email protected] for a pro-forma invoice. For university plans, credits are allocated per semester based on your agreement.
Can a university pay by invoice?
Yes. We issue a pro-forma and an invoice from KlusAI Labs SRL, with reverse charge for EU VAT-registered institutions.
Get started
Price it. Launch it.
It stops when you do. From €1,09/h excl. VAT, 60-min minimum.