Mistral AI: a European model family you can self-host
A European model vendor that will run in the customer’s environment. Strong for teams that want to self-host models; less so for companies that want a finished system.
Contents
| Type | European model vendor |
|---|---|
| Pricing | Published |
| Website | mistral.ai |
Mistral AI sells its models through an API and through Le Chat, its assistant. For sovereignty, the relevant offer is in its Team and Enterprise tiers: deployment self-hosted, in a private cloud or on-premises, so that data stays in the customer's environment1.
Best for: Companies that want a European model family they can run self-hosted, in a private cloud or on-premises, with a chat assistant for staff.
Not the right fit for: Companies that want a finished business system rather than models and an assistant to build on.
What Mistral AI offers on sovereignty
Its developer platform lists three deployment patterns: hybrid across cloud and on-premises, dedicated to isolate workloads, and self-hosted to "retain full control over infrastructure and data residency"2. Le Chat Enterprise adds private deployments with custom models, interface and tools1.
Prices are published for everything except Enterprise: Le Chat Pro at $14.99 a month, Team at $24.99 per user a month, and Mistral Large through the API at $0.5 per million input and $1.5 per million output tokens, with batch processing at half price1. The pricing page lists a training opt-out but does not make clear which plans it covers, and it does not say where the hosted service stores or processes data1.
Strengths and limits
Strengths
- Self-hosted deployment is offered for Team and Enterprise customers, with the stated aim that data stays in the customer’s environment.
- Published prices for the assistant and the API, including a 50% discount for batch processing.
- Hybrid, dedicated and self-hosted options on its developer platform.
Limits
- The pricing page does not say where data is stored or processed for the hosted service.
- Enterprise pricing and its training policy are not published.
- It supplies models and an assistant; business workflows still have to be built.
Choose another option if: You want AI to run operational work on your records without building it yourself, or you need model choice beyond one vendor.
Pricing
Published: Le Chat Pro $14.99 a month, Team $24.99 per user a month, Mistral Large API at $0.5 per million input and $1.5 per million output tokens. Enterprise on request.
Questions to ask Mistral AI
- Where is data processed for the hosted Le Chat and API services?
- Which plans are covered by the training opt-out, and is it on by default for business plans?
- What hardware does a self-hosted deployment of the models we need require?
- What does an Enterprise private deployment cost in the first year?
See how Mistral AI compares with the other options in the main guide.
Sources
- Pricing. Mistral AI. Accessed 8 Nov 2026.
- La Plateforme. Mistral AI. Accessed 8 Nov 2026.