Mistral AI rebranded Le Chat as Vibe to create a unified agent platform for productivity and software development. This platform includes Work Mode for multi-step tasks and Code Mode for software engineering. Work Mode connects to Google Workspace, Outlook, SharePoint, Slack, and GitHub to manage documents and communications. Code Mode runs sessions in isolated cloud sandboxes and allows users to run tasks in parallel. These sessions persist even when a user’s machine is turned off. The platform also includes a VS Code extension that brings the coding agent into the IDE with project context.
The unified Vibe platform replaces the previous fragmented landscape of Le Chat for general tasks and separate coding tools. This rebranding creates a single interface for both general enterprise productivity and specialized software development. The platform delivers features like task scheduling for automated workflows and a canvas editor for long-form document creation. Users can also use the Vibe CLI for terminal-native coding tasks. You should examine the deployment architecture before selecting a plan to ensure it meets your technical requirements.
Performance of Mistral Medium 3
Mistral Medium 3 powers Le Chat Enterprise and provides enterprise-grade capabilities for coding and multilingual comprehension. This model delivers performance comparable to Claude Sonnet 3.7. In benchmark tests, it performs at approximately 90% of the capabilities of Claude Sonnet 3.7. The model is suitable for business environments that require high speed and cost-effectiveness. It delivers responses at approximately 1,000 words per second through Flash Answers.
The Mistral model catalog contains several tiers for different workloads. Mistral Large 3 is the flagship model with 675 billion total parameters and 41 billion active parameters. Mistral Small 4 is a newer, more capable model with 119 billion total parameters and 6 billion active parameters. For users needing high throughput, Mistral Small 3.2 provides a cheaper alternative. Mistral Nemo is the most affordable model in the lineup with a price of $0.02 per million input tokens.
The context window for these flagship models is 256,000 tokens. This capacity allows for the processing of complex documents and large datasets. The platform also includes a code interpreter that executes Python in a sandboxed environment. This tool handles mathematical computations, data visualization with Matplotlib, and the processing of CSV or Excel files. Users can also use document OCR to extract text from uploaded PDFs or screenshots.
Privacy, sovereignty, and deployment options
Mistral AI provides the flexibility to deploy Le Chat Enterprise in a self-hosted environment, in a private cloud, or as a service within the Mistral cloud to ensure complete data control. This deployment flexibility allows organizations to choose an infrastructure that fits their specific security needs. The platform complies with the GDPR and the EU AI Act. Because Mistral is a French company, Le Chat stays under European law by default. This reduces the legal exposure to the US CLOUD Act that affects US-based providers.
The platform includes a mode to disable conversation history to protect sensitive information. Users can also opt-out to exclude their data from model training datasets. For the API, Mistral keeps data for 30 days for abuse monitoring before deleting it. Data used through active connectors in Le Chat does not train or fine-tune the models. The company holds ISO 27001 certification. SOC 2 Type II certification is in progress as of May 2026.
The Enterprise plan provides advanced administrative tools for organizational control. Admins manage user roles, product seats, and workspace memberships through the Admin Panel. They can also use SAML SSO to manage access through existing identity providers. This tier includes full audit logs to track all organizational activity. These tools ensure compliance with internal governance policies.
How will the integration of custom chip design affect the long-term pricing of inference?
Integration through MCP connectors
Le Chat Enterprise includes over 40 pre-built MCP connectors that connect to tools like Snowflake, Databricks, GitHub, and Slack to facilitate consistent data flow between different enterprise systems. These connectors allow the AI to search, summarize, and act across the entire organizational stack. The directory includes providers such as Asana, Atlassian, Box, Notion, Linear, and Sentry. These integrations allow engineering and data teams to pull context directly into their workflows.
Organizations can also use custom MCP connectors to link Le Chat to internal enterprise systems. This capability allows the AI to access proprietary data and tools that are not part of the standard connector directory. The platform supports connecting to Google Drive, SharePoint, and OneDrive to unlock intelligence from existing enterprise documents. Users can organize these external sources into complete knowledge bases.
The system allows for the creation of personal libraries of frequently used documents. Users can cite, extract, and analyze information from these libraries. This functionality works across both uploaded files and documents stored in Drive or SharePoint. The integration reduces the need for manual data movement between different software applications.
Automating tasks with no-code agents
The platform includes a built-in agent builder for creating custom AI workflows. These agents handle routine tasks such as invoice processing, meeting summarization, and expense reporting. Because the builder requires no code, non-technical users can create and deploy assistants that match specific business requirements. These agents connect to an organization’s apps and libraries for contextual understanding.
Custom agents can perform multi-step operations autonomously. This automation reduces the manual work required for complex workflows. The platform allows for the deployment of these agents anywhere, including on-premises or in a private cloud. This ensures that automation does not bypass established security boundaries.
Users can also use the task scheduling feature to run prompts, agents, or workflows at set times. This allows for the automation of recurring business processes. The implementation of custom agents provides a way to scale expertise across a team without requiring constant human intervention.
Separate billing for Le Chat and the API
Mistral AI uses two separate billing systems for its products. Le Chat is a subscription model with monthly tiers for different user needs. The Mistral API is a pay-per-use service through la Plateforme. The API bills users per million tokens processed, with separate rates for input and output. A Le Chat Pro subscription does not cover API calls, and API credits do not apply to Le Chat.
The pricing for Le Chat varies by the selected plan. The Free tier is $0 per month and includes 25 messages per day. The Pro tier costs $14.99 per month and includes Mistral Vibe and a no-telemetry mode. The Team tier costs $24.99 per user per month, or $19.99 per user per month if billed annually. Enterprise pricing is custom and requires a direct quote.
| Plan | Monthly Price | Key Feature |
|---|---|---|
| Free | $0 | 25 messages per day |
| Pro | $14.99 | No Telemetry Mode |
| Team | $19.99 (Annual) | 30GB storage per user |
| Enterprise | Custom | On-premises deployment |
The entry price for genuine enterprise deployments starts around $20,000 per month, which makes it ten times more expensive than mid-market alternatives. For developers, the API provides access to models like Mistral Large 3 at $0.50 per million input tokens and $1.50 per million output tokens. Mistral also provides a 90% discount on repeated prompt prefixes through prompt caching.
Managing users and permissions
The Admin Panel allows organization admins to manage all members and their access levels. The Members page lists active users and invited users. Admins can change user roles, which include Admin, Billing, and Member. They can also toggle Vibe seats or Mistral Code seats for specific users.
The platform uses SAML SSO to manage user access through an identity provider. When SAML SSO is enabled, admins can turn on automatic seat assignment for users. This means users receive a seat when they sign in through the configured identity provider. Users in the Team and Enterprise plans can also use the Admin API to manage users and permissions programmatically.
Workspace roles are separate from organization roles. A member with a Billing role at the organization level can hold a Workspace Contributor role within a specific workspace. Admins use the Workspaces tab to add, remove, or change roles for members within individual project folders. This structure allows for granular control over who can access specific documents and conversations.
Enterprise suitability and final verdict
Le Chat Enterprise is designed for large organizations with over 1,000 users and developer-heavy workflows. The 40 plus MCP connectors and the ability to deploy on-premises make it a strong choice for companies with strict data residency requirements. It fits organizations in regulated sectors like finance, healthcare, and government that need to maintain data sovereignty. The platform is a good fit for companies that already use Google Workspace or Microsoft 365.
The platform is not a good fit for mid-sized organizations with fewer than 1,000 users. Smaller companies may find mid-market alternatives like Teamo AI or LangDock to be significantly cheaper. The product also lacks native messaging integrations for platforms like WhatsApp, Signal, or Teams. Organizations with a large non-desk workforce may find the web and mobile app interface insufficient for their needs.
Le Chat Enterprise is a premium product for large-scale, regulated enterprises. It provides a way to centralize AI productivity while maintaining strict control over data and deployment. Organizations should prioritize their data governance and integration needs before committing to a custom enterprise contract.
