Pricing

You pay only forwhat you turn on.

For Savings you pay 20% of the savings that actually happen. If no savings arise, there is no savings invoice. Shield and Sovereignty are switches on the same API, and you pay only for those that are on.

10-day pilot for free, you keep all the savings.
Your data trains only your own savings model.

Calculate your own savings
€/mo
Savings from routing, 18%€900/mo
AIvantage’s share, 20% of savings€180/mo
You keep€720/mo
Prices

Three switches, three prices.

You turn on the ones you need. The bill for cloud models keeps coming straight from the model provider, because you use your own OpenAI, Anthropic, Google or xAI key. Prices exclude VAT.

Savings
20%of actual savings
No savings, no savings fee.
i
Savings are calculated by comparing the actual price with what the same request would have cost on the most expensive model you allow, at public list price, request by request. You get the request log as a CSV file. On locked-model requests you pay 20% only of what AIvantage’s cache saved.
Read more about Savings
Shield
€29/moper API key
Up to 20,000 queries per month per key.
i
During the 10-day pilot Shield only shows its findings, and it is not billed.
Read more about Shield
Sovereignty
€0.20from / 1M input tokens (excl. VAT)
Fixed-price from €1,590/mo.
i
Output tokens cost four times the input price.
Current selection, € / 1M tokensinputoutput
Qwen 2.5 3B0.200.80
Qwen 2.5 7B0.401.60
Qwen 2.5 32B1.405.60
Qwen 3.6 27B2.409.60
Read more about Sovereignty
Support
Basic supportIncludedWe reply as best we can by the next business day.
Priority support€120/moWe reply as quickly as possible. Support work €120/h, each started 15 minutes per support request per month.
Other work€120/hFor example an audit, per started hour.
Questions

Pricing as questions

The pilot lasts 10 days and is free. You keep all the savings, and savings are on from the first request. The key and the customer relationship stay yours throughout, and the model provider’s bill comes as before.

After the pilot you get a summary of the savings, Shield’s findings and the quality spot checks, and then you decide whether to continue. You can stop at any time by deleting the key in the portal.

The reference price is the public list price per million tokens of the most expensive model you allow. It is preset for you, so getting started requires nothing, and you can change it in the portal. Savings are the difference between the reference price and the actual price for exactly the same requests. The request log shows both prices for every request, and you can download the log as a CSV file and calculate it yourself.

Then savings are not billed. On locked-model requests only the saving from caching is billed, and if a period ends in the negative, the savings share is €0. The report shows the negative as it is.

The model provider, as before. You use your own OpenAI, Anthropic, Google or xAI key, so the bill for cloud models comes straight from the provider. From AIvantage you receive only the share for the switches that are on.

After the pilot you get a summary and decide whether to continue. If you don’t activate the service, the API stops when the pilot ends, and nothing is billed for the pilot. You receive an email summary of the pilot savings. If you continue, you can stop at any time by deleting the key in the portal.

Savings apply to traffic through the API, meaning applications, integrations and coding assistants that use an API key. AIvantage does not connect to per-user licences such as ChatGPT or Copilot subscriptions. If there is no API traffic, there is nothing to save, and no invoice either.

AIvantage
AI from the plug

The savings of the first 10 days are all yours.

Get started

10-day pilot for free, you keep all the savings.
Your data trains only your own savings model.

janne@aivantage.fi/+358 40 577 0215