Deployment and Pricing

Update November 2, 2026 — Le Chat is now Vibe. As of May 28, 2026, Le Chat has been rebranded as Vibe (still at chat.mistral.ai): a single agent for work and code that carries over your conversations, settings, and subscriptions. The legacy Le Chat experience remains available in Chat mode; in Work mode, agents are now skills (invoked with @), memories form the knowledge base, and deep research is the Deep Research skill. Plans are now called Mistral Free, Pro, Team, and Enterprise. This lesson uses the name Vibe; note that some menus described here may have changed names or locations.

Three plans, three organizational realities

Mistral AI offers Vibe in three plans. The difference between them isn’t response quality—it’s the level of governance and support surrounding the assistant. The Pro plan targets independent professionals and very small teams. It unlocks access to all Mistral models, including the latest versions, plus web search, document analysis, Canvas, and code generation. You get priority queuing and email support. A consultant preparing client deliverables finds everything needed here, precisely because they don’t need to connect to a corporate directory or produce audit logs.

The Team plan introduces collaboration for groups of five to two hundred people. It includes everything in Pro and adds a shared team space where conversations and agents circulate among members. You get centralized management assigned to an administrator, unified billing on a single invoice, the ability to create agents accessible to everyone, and high-priority queuing. This is the natural entry point for a marketing department or legal team that wants to structure its AI usage without launching a full-scale enterprise project.

The Enterprise plan builds on Team and adds the full governance stack covered in the previous lesson: SSO/SAML 2.0 with automatic provisioning, contractually guaranteed zero retention, detailed and exportable audit logs, an admin console with granular policies, access to Enterprise Search, dedicated support with a guaranteed SLA, a dedicated account manager, and model customization on your data via Forge. Pricing is quote-based, calibrated to user volume and activated services. Bring a seat count estimate to your first sales meeting: that variable structures the entire discussion.

Where to run the platform

Hosting decisions rarely hinge on technical criteria. They stem from the regulatory constraints your organization faces.

Mistral Cloud remains the default option and suits the vast majority of companies. Data lives on Mistral’s infrastructure in Europe, with logical isolation by organization. The new Paris data center, funded by the $830 million raise in March 2026, strengthens this sovereign hosting capacity. Dedicated VPC is the next step up, designed for organizations that require complete network isolation. Mistral still manages the infrastructure, but it sits physically separate from other customers’ environments. This gives a bank or insurer a defensible position when talking to regulators.

On-premise deployment serves sectors where data must never leave the organization’s perimeter—defense, intelligence, nuclear. Mistral deploys directly on customer servers as a software license accompanied by tailored technical support. Finally, Partner Cloud relies on Mistral models available on Microsoft Azure, Amazon Bedrock, and Google Cloud. This is the shortest path for an organization already committed to one of these ecosystems, since existing billing and access controls apply without negotiating a new master agreement.

How an Enterprise deployment unfolds

An Enterprise project typically spans two to four months across four phases. Qualification takes one to two weeks: you assess needs with a Mistral sales rep, define scope by user count and priority use cases, lock in the deployment option, and receive the commercial proposal. Technical setup follows over two to four weeks: configure SSO with your identity provider, enable the admin console, create initial roles and policies, run integration tests, and validate security.

The third phase, gradual rollout, runs four to eight weeks and represents the real moment of truth. It starts with a pilot among a restricted group of 50 to 100 users, includes training for administrators and champions, then collects feedback to drive adjustments before extending to the whole organization. Champion selection deserves far more attention than it usually gets. These are the people who will answer colleagues’ questions, long before your internal support team gets involved. Continuous optimization then takes over with no end date: analyze adoption metrics, add use cases and specialized agents, hold quarterly reviews with your Mistral account manager, and evolve policies based on observed usage.

Contractual support follows the same duration logic. The Enterprise offer includes technical support with a guaranteed SLA where response time varies by severity, a dedicated account manager for strategic follow-up, initial training for administrators and key users, private technical documentation with integration guides, and a dedicated Slack or Teams channel for daily exchanges. That last point might seem minor, but it avoids the classic trap of a ticket opened Friday night and handled the following Tuesday.

Key takeaways

• Three plans (Pro, Team, Enterprise) cover every need, from individual users to large accounts
• Four deployment options meet even the strictest sovereignty requirements
• Enterprise onboarding follows a structured four-phase process over 2 to 4 months
• Dedicated support with a guaranteed SLA accompanies organizations long-term
• Enterprise pricing is quote-based, adapted to volume and activated services