AI platform: the layer
above the card
AI platform from data centres in Germany and Finland: notebook, training run, endpoint and ready-to-use models (including LLM hosting) โ as a service instead of an empty machine, 6 open-weight models in the catalogue. Billed per second of runtime or per million tokens. A training run trains a model on your data, an endpoint is the fixed network address where a ready-to-use model responds; the underlying hardware is the same as that rented by the hour in the GPU catalogue โ you work on the model, not on the machine.
Prices are net plus 19% VAT., location Frankfurt. If you want the card without the platform on top, you can find it under GPU.
- Accelerator
- โฌ0.7397$0.8581 to โฌ5.2500$6.0900 per card and hour
- Workspaces
- Jupyter and VS Code
- Training runs
- 1 to 8 nodes
- Endpoints
- Scale to zero
- Model catalogue
- 6 models, open weights
- Billing
- per second or per token
- Inference locations
- 3 in Germany and Finland
- Accelerators from
- โฌ0.7397$0.8581
- per card and hour, L4 with 24 GB
- Workspace from
- โฌ0.0567$0.0658
- per hour, billed by the second
- Output from
- โฌ0.1251$0.1451
- per million tokens โ a token is roughly a part of a word
- Inference locations
- 3
- Inference means: a finished model responds. No processing outside the EU.
Three layers, three billing units
Accelerator, service and model are not the same product. They differ in what you need to bring and how they are billed.
Layer 1 ยท Hardware
Rent accelerators
Billing per hour, per card
You get an instance with one, two, four or eight cards, root access and an empty machine with drivers. Operating system, framework, model state and scaling are up to you. You can find the hourly rate further down this page.
Right for you if you have your own toolchain and need full control over the environment.
Layer 2 ยท Platform
Notebook, training run, endpoint
Billing per service runtime
The same hardware, but as a service: Jupyter starts ready to use, a training run is queued instead of logged in, a model becomes an HTTPS endpoint with scale-to-zero.
Right for you if you work on models and not on machines.
Layer 3 ยท Model
Call ready-to-use models
Billing per million tokens
No node, no deployment, no waiting time. A call against a model name; you are billed for what went through the wire.
Right for you if you are building an application and do not want to operate a model.
What an hour of computing costs
Five accelerator classes support every row on this page where a card appears โ from the accelerated workspace to the endpoint. The amounts are taken unchanged from the instance catalogue. Only the reference unit is new here: the rate per node is broken down to the individual card and to the hour.
What a card can support is determined by its graphics memory โ the VRAM, the memory on the card itself. It ranges from 24 GB on the L4 to 141 GB on the H200 SXM. A model whose weights do not fit will not run on this card โ which is why the memory size is in the same row as the price.

| Class | Graphics memory | Cards per node | Card / h | Node / h | Node / month |
|---|---|---|---|---|---|
| NG-L4Inference of small and quantised models, embeddings | 24 GB per card | 1 | โฌ0.7397$0.8581 | โฌ0.7397$0.8581 | โฌ540.00$626.40 |
| NG-L40SModels up to around 32 billion parameters, image models | 48 GB per card | 1 | โฌ1.3808$1.6017 | โฌ1.3808$1.6017 | โฌ1,008.00$1,169.28 |
| NG-RTX6000Mixture of experts models and long contexts โ 96 GB VRAM | 96 GB per card | 1 | โฌ1.6425$1.9053 | โฌ1.6425$1.9053 | โฌ1,199.00$1,390.84 |
| NG-H100Training runs โ the reference class for all calculations below | 80 GB per card | 1 | โฌ2.6575$3.0827 | โฌ2.6575$3.0827 | โฌ1,940.00$2,250.40 |
| NG-H200ร8Large models across multiple cards, distributed training | 141 GB per card | 8only as a whole node | โฌ5.2500$6.0900 | โฌ42.0000$48.7200 | โฌ30,660.00$35,565.60 |
All amounts net, plus 19% VAT., Frankfurt location. The monthly rate is maintained; the hourly rate is this amount divided by 730 hours. The four single-card classes can be ordered individually; the eight-card class is only allocated as a whole node, so its card rate is a calculated value and not a separate quote.
The platform either adds a named surcharge to these rates โ for the workspaces โ or converts them into a token price via a measured throughput. A token is the building block into which a model breaks down text โ usually a short word or a syllable. Both calculation methods are shown below the tables in the following sections.
A notebook that is already running
A workspace is an instance with a ready-to-use interface: Jupyter and VS Code are set up, the drivers are correct, and the Object Storage is mounted as a directory. This is where fine-tuning happens โ retraining a pre-trained model on your own data. Runtime is billed, not ownership.
- Jupyter and VS Code
- Both interfaces run on the same environment and can be switched during operation. VS Code is also accessible via the desktop editor's remote connection.
- Accelerators can be added
- The environment starts without a GPU and can be upgraded to an accelerated class at the next start. The working directory is retained, the price changes with the class.
- Billing by runtime
- The time the environment is running is counted โ to the second, not the time it is created. A stopped environment only costs its storage.
- Data connection to Object Storage
- Buckets are mounted as a directory at startup. Access keys are stored in the project, not in the notebook; they do not appear in the cell history or in checkpoints.
What a workspace costs
The hourly rate of a workspace is the hourly rate of its instance class plus a 15% platform fee. This fee covers image maintenance, operating both interfaces, idle detection, and the S3 connection, i.e. access to Object Storage. The instance price itself remains the same as in the catalogue โ the platform does not make the card more expensive, it adds a specified amount alongside it.
| Class | Configuration | Accelerator | Instance / h | Platform / h | Environment / h | Cap / month |
|---|---|---|---|---|---|---|
| NP3Data preparation, analysis, classic machine learning without accelerators | 4 vCPU ยท 8 GB RAM | without accelerator | โฌ0.0493$0.0572 | โฌ0.0074$0.0086 | โฌ0.0567$0.0658 | โฌ41.39$48.01 |
| ND3The same work with constant performance โ no shared CPU time | 4 vCPU ยท 16 GB RAM | without accelerator | โฌ0.1063$0.1233 | โฌ0.0159$0.0184 | โฌ0.1222$0.1418 | โฌ89.21$103.48 |
| NG-L4First experiments with quantised language models, embeddings, image models | 8 vCPU ยท 48 GB RAM | 1 ร L4 | โฌ0.7397$0.8581 | โฌ0.1110$0.1288 | โฌ0.8507$0.9868 | โฌ621.01$720.37 |
| NG-L40SLoRA adaptations and models up to around 30 billion parameters | 16 vCPU ยท 96 GB RAM | 1 ร L40S | โฌ1.3808$1.6017 | โฌ0.2071$0.2402 | โฌ1.5879$1.8420 | โฌ1,159.17$1,344.64 |
| NG-H100Full fine-tuning of medium-sized models, HPC workloads with FP64 | 32 vCPU ยท 256 GB RAM | 1 ร H100 SXM | โฌ2.6575$3.0827 | โฌ0.3986$0.4624 | โฌ3.0561$3.5451 | โฌ2,230.95$2,587.90 |
All amounts net, plus 19% VAT., Frankfurt location. The monthly cap corresponds to 730 hours: a continuously running workspace never costs more, regardless of the number of calendar days.

A run is billed by runtime at ENTRONYX CLOUD, not by result โ even a cancelled one costs the seconds it ran. However, it resumes from the last complete checkpoint and not from the beginning.
A run is a job, not a session
You do not log in to a machine, but submit a job. The platform finds the nodes, mounts the data, writes checkpoints โ secured intermediate states from which a run can resume โ and releases the cards again as soon as the last step has run.
5 steps
Job flow
- 0101SubmitA job consists of an image, command, data source and the desired class. It is submitted via the console, the API or from a pipeline โ a session is not required for this.
- 0202QueueIf the class is occupied at the selected location, the job waits instead of failing. The estimated start time is shown in the job; jobs in a project queue run in order.
- 0303ExecuteAccelerators, storage and network are allocated, the image is started, the data source is mounted. Across multiple nodes, the platform sets the rank and the addresses of the peers itself.
- 0404CheckpointsA checkpoint directory is written to Object Storage. An aborted job resumes at the last complete checkpoint, not at step zero.
- 0505Log and endStandard output, metrics and events remain accessible for 30 days and can be downloaded as a file. With the last step, the nodes are released and billing ends.
Which model size runs on how many cards
The weights of a model must fit into the graphics memory. At 8 bits โ quantisation, i.e. shortening each number to one byte โ one billion parameters occupy around one gigabyte; at BF16 with two bytes per number, it is double. Therefore, a card with 24 GB is sufficient for a model with eight billion parameters, while a model with seventy billion needs two cards. The rows below are the allocation with which the model catalogue on this page actually runs.
| Model | Parameters | Number format | Weights | Cards per replica | Graphics memory |
|---|---|---|---|---|---|
| Ministral 8B Instruct | 8 billion | 8 bit | 8 GB | 1 ร L4 | 24 GB |
| Mistral Small Instruct | 24 billion | 8 bit | 24 GB | 1 ร L40S | 48 GB |
| Qwen2.5 Coder 32B | 32 billion | 8 bit | 32 GB | 1 ร L40S | 48 GB |
| Mixtral 8ร7B Instruct | 47 billion, 13 billion active | 8 bit | 47 GB | 1 ร RTX PRO 6000 Blackwell Max-Q | 96 GB |
| Llama 3.3 70B Instruct | 70 billion | 8 bit | 70 GB | 2 ร H200 SXM | 282 GB |
| BGE-M3 | 568 million | BF16 | 1.2 GB | 1 ร L4 | 24 GB |
A replica is a running copy of the model. The graphics memory must hold the cache of running requests in addition to the weights โ which is why the column on the right is consistently larger than the weights column and not tightly sized.
What distribution costs
The reference is a run over 72 hours on a node of class NG-H100 at โฌ2.6575$3.0827 per hour โ the same calculation as on the GPU page, just distributed across multiple nodes here. More nodes shorten the wait time, but they do not lower the price: gradient exchange โ the synchronisation of learning steps between nodes โ costs efficiency, and you pay for GPU hours.
| Nodes | Efficiency | Elapsed time | GPU hours | Costs of the run |
|---|---|---|---|---|
| 1 ร NG-H100One node, no network traffic between cards | 100% | 72.0 h | 72.0 h | โฌ191.34$221.95 |
| 2 ร NG-H100Gradient exchange via the RDMA fabric in the vRack, the private network between the nodes | 94% | 38.3 h | 76.6 h | โฌ203.55$236.12 |
| 4 ร NG-H100Gradient exchange via the RDMA fabric in the vRack, the private network between the nodes | 88% | 20.5 h | 81.8 h | โฌ217.43$252.22 |
| 8 ร NG-H100Gradient exchange via the RDMA fabric in the vRack, the private network between the nodes | 80% | 11.3 h | 90.0 h | โฌ239.18$277.45 |
Efficiency is the share of theoretical acceleration that remains after deducting gradient exchange. The costs are calculated as GPU hours times โฌ2.6575$3.0827. All amounts net, plus 19% VAT.
The row with eight nodes reads like a trade-off: the run finishes after 11.3 instead of 72 hours, but costs โฌ47.84$55.49 more than on one node. Whether this is worth it is not decided by the price, but by the question of how often you want to see a result in a day.
How long your own run takes depends on data volume and method; the catalogue does not answer this. It answers the price: you pay for the occupied GPU hours, billed to the second. A run that aborts halfway through costs half โ there is neither a started day nor a minimum term.

From checkpoint to HTTPS endpoint
You provide weights or an image, and the platform turns them into an address with an access key, certificate and scaling. You are billed for the time a replica โ a running copy of the model โ is running, not the time the endpoint is deployed.
Scale to zero
Without requests, the endpoint scales down its replicas and costs nothing. The first request afterwards triggers a cold start โ loading the weights into the GPU memory before the first response arrives. All subsequent requests hit a running replica. If you want to avoid this waiting time, set the lower limit to one โ scaling up is then elastic and scaling down is limited. If you prefer to manage the cards in your own cluster, book GPU nodes in Managed Kubernetes.
At least zero replicas
The first request after an idle period waits for the cold start
Internal tools, batch processes, anything used during office hours
At least one replica
No cold start time, one replica runs continuously
Applications with users waiting for a response
At least two replicas
The endpoint responds even if a fire zone fails
Services with an availability commitment
Calculation example
An endpoint that only responds during the day
One replica on NG-L4, active for 8 hours per day over 30 days. Outside this time, the endpoint is at zero.
- Class hourly rate
- โฌ0.7397$0.8581
- Active hours per month
- 240 h
- Cost with scaling to zero
- โฌ177.53$205.93
- Same class continuously
- โฌ540.00$626.40
- Savings compared to continuously
- 67%
Net amounts, Frankfurt location. The hourly rate is the same as in the instance catalogue โ the platform just bills it to the second and does not charge for time without a replica.
Cold start and concurrent requests
A cold start consists of two parts: the weights are read from the local NVMe at 6 GB/s, then the runtime needs 11 seconds for CUDA context, graph construction and the first health check. Both values are in the catalogue file, so the column below can be verified and is not just a claim.
| Model in the endpoint | Precision | Weights | Class | Cold start | Concurrent requests |
|---|---|---|---|---|---|
| Language model, 8 billion parameters | 8 bit | 8 GB | NG-L4 | 12.3 s | 24 |
| Language model, 24 billion parameters | 8 bit | 24 GB | NG-L40S | 15.0 s | 32 |
| Image model, diffusion | BF16 | 7 GB | NG-L40S | 12.2 s | 12 |
| Language model, 70 billion parameters | 8 bit | 70 GB | NG-H100 | 22.7 s | 8 |
| Embedding model, 568 million parameters | BF16 | 1.2 GB | NG-L4 | 11.2 s | 64 |
Concurrent requests apply per replica with 4,096 tokens of context and an 8-bit cache; a longer context lowers the value. Requests above this limit start another replica instead of failing.
You are billed for what went over the wire
A call to a model name, with no node and no deployment. The catalogue contains only models with open weights, and they run on our own cards at locations in Germany and Finland.
The context column states the context window: how many tokens โ question and previous answer combined โ a model can oversee at the same time. Anything beyond that falls off the end, which is why the number is a limit and not an advertising value.
| Model | Task | Context | Class | Throughput | Input / million | Output / million |
|---|---|---|---|---|---|---|
| Ministral 8B Instruct8 billion ยท Apache 2.0 | Dialogue and tool calls | 32,768 tokens | NG-L4 | 912 tokens/s | โฌ0.0704$0.0817 | โฌ0.2816$0.3267 |
| Mistral Small Instruct24 billion ยท Apache 2.0 | Dialogue and tool calls | 128,000 tokens | NG-L40S | 1,152 tokens/s | โฌ0.1041$0.1208 | โฌ0.4162$0.4828 |
| Qwen2.5 Coder 32B32 billion ยท Apache 2.0 | Program code | 32,768 tokens | NG-L40S | 648 tokens/s | โฌ0.1850$0.2146 | โฌ0.7399$0.8583 |
| Mixtral 8ร7B Instruct47 billion, 13 billion active ยท Apache 2.0 | Dialogue and tool calls | 32,768 tokens | NG-RTX6000 | 4,560 tokens/s | โฌ0.0313$0.0363 | โฌ0.1251$0.1451 |
| Llama 3.3 70B Instruct70 billion ยท Llama 3.3 Community | Dialogue and tool calls | 128,000 tokens | NG-H200ร8 | 4,224 tokens/s | โฌ0.2158$0.2503 | โฌ0.8631$1.0012 |
| BGE-M3568 million ยท MIT | Embeddings | 8,192 tokens | NG-L4 | 179,200 tokens/s | โฌ0.0014$0.0016 | no output price |
All amounts net, plus 19% VAT. The throughput is the aggregated value of a replica across all concurrent streams, measured with continuous batching. Embedding models return vectors โ series of numbers used to compare texts โ and therefore have no output price.
How the token price is calculated
We did not invent a token price, but divided the hourly rate of the card by the throughput. A replica costs an amount from the instance catalogue per hour and delivers a measurable number of tokens in that hour. The quotient is the cost rate per token; divided by the assumed utilisation, it results in the list price.
Input tokens are cheaper because in the prefill โ reading the request โ they run across the tensor cores and are not bound by memory bandwidth like decoding, generating the response token by token. It is set at 25% of the output price.
Why the prices vary so much between models is shown in the throughput column: a model with mixed experts โ Mixture of Experts in English โ activates only a fraction of its weights per token and is therefore cheaper than a dense model of the same size. It is not the parameter count that determines the price, but the number of tokens per card hour.
Calculation basis
- Hourly rate
- from the instance catalogue, 730-hour month
- Throughput
- aggregated per replica, all streams
- Assumed utilisation
- 80%
- input to output
- 25%
- utilisation in batch mode
- 95%
- throughput in batch mode
- ร 1.6
The assumed utilisation is below one hundred percent because an endpoint reserves capacity for load peaks. This reserve costs us money and is therefore included in the price, instead of being passed on to the caller as waiting time.
Batch processing
If you do not need to guarantee a response time, you submit your requests as a file and collect the result as soon as it is ready โ usually within 24 hours. Without a latency commitment, we can pack more densely: utilisation increases from 80% to 95%, and larger batches raise the throughput per card by a factor of 1.6. Both together result in the discount of 47% โ not a negotiated discount, but the amount we actually save.
| Model | Synchronous input | Synchronous output | Batch input | Batch output |
|---|---|---|---|---|
| Ministral 8B Instruct | โฌ0.0704$0.0817 | โฌ0.2816$0.3267 | โฌ0.0371$0.0430 | โฌ0.1482$0.1719 |
| Mistral Small Instruct | โฌ0.1041$0.1208 | โฌ0.4162$0.4828 | โฌ0.0548$0.0636 | โฌ0.2190$0.2540 |
| Qwen2.5 Coder 32B | โฌ0.1850$0.2146 | โฌ0.7399$0.8583 | โฌ0.0974$0.1130 | โฌ0.3894$0.4517 |
| Mixtral 8ร7B Instruct | โฌ0.0313$0.0363 | โฌ0.1251$0.1451 | โฌ0.0165$0.0191 | โฌ0.0658$0.0763 |
| Llama 3.3 70B Instruct | โฌ0.2158$0.2503 | โฌ0.8631$1.0012 | โฌ0.1136$0.1318 | โฌ0.4542$0.5269 |
| BGE-M3 | โฌ0.0014$0.0016 | no output price | โฌ0.0007$0.0008 | no output price |
Amounts net. A batch job is accepted as soon as it is submitted; processing begins when capacity becomes available. Results are stored as a file in the Object Storage of the project.

A cancelled run resumes from the last complete checkpoint, not at step zero: lost are the hours since this checkpoint, not the 72 of the entire run. The checkpoints are located in the Object Storage of the same location โ at ENTRONYX CLOUD, neither a checkpoint nor a request leaves the EU.
Reference run 72 h ยท 11.3 h on 8 nodes
Where inference runs and what happens to your inputs
With an AI service, the operating location is not a footnote. Anyone who sends text to a model hands it over. The short answer: ENTRONYX CLOUD does not train on customer data, and no input leaves the EU. There is no path to a third country because no sub-processor is involved in inference. The long answer is below in sentences, not in a list of seals.
| Question | Answer |
|---|---|
| Where does the inference run? | Exclusively at the locations in Germany and Finland where the platform runs. A call without a location specified goes to the nearest one, never beyond. An outflow to a third country โ any country outside the EU and the EEA โ is therefore excluded, and outsourcing to a model provider does not take place. |
| What happens to the inputs? | They are kept in memory for the duration of the request and then discarded. No writing to storage, no storing in a cache, no training on your data โ neither for our models nor for those of third parties. This commitment is also stated in the data processing agreement. |
| What is logged? | Time, model name, token count, response time and status code โ the details from which the invoice is generated. Not the content of the request and not the content of the response. |
| Is there any transfer to third parties? | No. There is no sub-processor for inference. The models are stored as open weights on our own storage media and run on our own hardware. |
| Who is the legal operator of the platform? | A German company based in Germany, with no parent company outside the EU. Requests for disclosure under foreign law are therefore void; the mutual legal assistance procedure is authoritative. |
| What is in the data processing agreement? | The agreement under Article 28 GDPR names the locations individually, excludes processing outside the EU and lists the technical and organisational measures. It can be viewed before the contract is concluded. |
| Are the models verifiable? | The weights are open and their licence is noted in the catalogue. The model version and checksum of each endpoint are in the console; a model change gets a new name and never silently replaces the old one. |
These details are part of the data processing agreement. If the contract text deviates from this table, the contract applies โ please report the discrepancy to us, as it is then an error on this page.
Inference locations
3 locations in Germany and Finland. The complete model catalogue is served at 3 of them; the others run the models whose instance class is listed there. A request without a location specified will always remain within this list. PUE is the energy efficiency metric of a data centre โ the total power consumption divided by the power of the computing equipment; 1.0 would be lossless.
Frankfurt am Mainfra1
Location Rhein-Main ยท Germany
PUE 1.14 ยท 3 ms
- ISO 27001
- ISO 50001
- BSI C5:2020
- EN 50600 VK4
- TรV Rheinland
Complete model catalogue
Dรผsseldorfdus1
Location Rhein-Ruhr ยท Germany
PUE 1.12 ยท 4 ms
- ISO 27001
- BSI C5:2020
- EN 50600 VK4
Complete model catalogue
Helsinkihel1
Location Uusimaa ยท Finland
PUE 1.08 ยท 21 ms
- ISO 27001
- ISO 50001
- BSI C5:2020
- EN 50600 VK4
Complete model catalogue
What we do not offer
- No passing on to third-party model APIs
- We do not operate a proxy system that forwards requests to an American provider and passes them off as our own offering. What is in the catalogue runs on our cards.
- No closed weights
- We cannot run a model on our own hardware if we are not allowed to deliver its weights. Therefore, the catalogue only contains models with an open licence.
- No silent model changes
- A model name remains tied to a weight version. New versions appear as a separate name; old ones are phased out with an announced notice period.
What is usually asked before the first AI service
Ten questions that regularly come up before ordering. The answers use the same catalogue values as the sections above.
Your question is not included? The AI configurator guides you through the workspace, training job and endpoint, with the same hourly rates as the tables on this page.
What does LLM hosting in Germany and Finland cost, and which models are running?
LLM hosting at ENTRONYX CLOUD means: Six ready-to-use language and embedding models with open weights run on our own cards in data centres in Germany and Finland and are billed per million tokens, with no base fee; the output price of the cheapest model is shown in the line below this answer. You operate your own models as an endpoint based on runtime, scaling to zero when no one is calling them.
Output from, per million tokens, netโฌ0.1251$0.1451
What differentiates the AI platform from a GPU instance?
With the GPU instance, you rent the card per hour and operate the operating system, framework and scaling yourself. The platform turns the same hardware into a service: workspace, training job and endpoint are billed according to their term, ready-to-use models per million tokens.
Accelerators from, per card and hour, netโฌ0.7397$0.8581
How much does a workspace with Jupyter cost?
The hourly rate of its instance class plus 15% platform share, billed to the second based on the term. Without cell execution and without an open connection, it stops automatically after 60 minutes; when stopped, it only costs its storage.
Cheapest environment, NP3, per hour netโฌ0.0567$0.0658
How is a training run billed?
By occupied GPU hours, billed to the second, without started days and without a minimum term. Distributed across up to eight nodes, a run finishes earlier, but is not cheaper: the gradient exchange between the nodes costs efficiency, and you pay for GPU hours.
Reference run, 72 hours on one node NG-H100, netโฌ191.34$221.95
What happens if a training run aborts?
It resumes at the last complete checkpoint, not at step zero; the checkpoints are stored in the Object Storage of the same location. You pay for the seconds the run was active, because billing is based on the term, not the result.
Does an endpoint cost money even if no one calls it?
Not with scaling to zero: without requests, the endpoint shuts down its replicas and costs nothing, the first request afterwards waits for the cold start. A replica on NG-L4, active 8 hours a day over 30 days, thus costs 67% less than the same class continuously.
Calculation example with scaling to zero, per month netโฌ177.53$205.93
Which ready-to-use models are available, and how are they billed?
Six open-weight models: Ministral 8B Instruct, Mistral Small Instruct, Qwen2.5 Coder 32B, Mixtral 8ร7B Instruct, Llama 3.3 70B Instruct and BGE-M3. Billed per million tokens, for text models input tokens at 25% of the output price; failed requests do not count. Without a response time commitment, batch processing costs 47% less.
Output from, per million tokens, netโฌ0.1251$0.1451
At which locations does the AI platform run?
At three locations in Germany and Finland: Frankfurt am Main (fra1), Dรผsseldorf (dus1) and Helsinki (hel1). The complete model catalogue runs on each of them. A request without a specified location will always remain within this list.
How fast does support respond, and how do I reach them?
Via ticket in the account: The price includes the basic support of the SLA with an initial response in 8 hours for critical incidents (P1), available Monday to Friday from 08:00 to 18:00. Initial response means: A human has accepted the report, classified it and reported back the processing status โ the SLA does not guarantee a resolution time. In addition to basic support, it offers three further support levels; the highest responds to P1 in 15 minutes.
Is there a minimum term, and how do I cancel?
Not without a term commitment: You can delete services billed by the hour at any time without notice in the console or via the API, and billing ends at the time of deletion (Section 11 paragraph 3 of the Terms and Conditions). You can cancel monthly contracts with one day's notice to the end of the term, contracts for 12 or 24 months with 30 calendar days โ otherwise they renew for twelve months (Section 11 paragraph 2 of the Terms and Conditions). You pay the term in advance and receive a 5% or 8% term discount in return.
Configure a service or rent the card directly
The configurator guides you through the working environment, training job and endpoint, showing the same hourly rate as in the tables above. If you need the hardware without the platform on top, go directly to the GPU catalogue; the complete overview of all items is in the price list.
- Accelerators from
- โฌ0.7397$0.8581
- Workspace from
- โฌ0.0567$0.0658
- Output from
- โฌ0.1251$0.1451
- Inference locations
- 3
