Mistral Le Chat: Features, Pricing, Models & Alternatives (2026)

Last researched: September 26, 2026. Pricing and plan details change frequently — always confirm current figures at mistral.ai/pricing before purchasing. Mistral Le Chat is the consumer- and business-facing AI assistant built by Mistral AI, the French AI lab behind the open-weight Mistral model family. In 2026, Mistral rebranded Le Chat as "Vibe," folding the original chat assistant into a broader agent that spans everyday productivity work and software development, while keeping the same account, login, and URL (chat.mistral.ai). This profile covers the product under both names since "Le Chat" remains the name most people search for.

What Is Mistral Le Chat?

Mistral Le Chat launched as Mistral AI's answer to ChatGPT and Gemini: a web and mobile chat interface for asking questions, summarizing and analyzing documents, generating images, browsing the web, and writing code, all backed by Mistral's own large language models. In 2026 Mistral restructured the product into "Vibe," an agent with three modes — Vibe Chat (the original turn-based conversation experience), Vibe Work (delegating multi-step productivity tasks across connected apps), and Vibe Code (an agentic coding assistant for the CLI, IDEs, and web). Functionally it is still the same assistant users knew as Le Chat, expanded with more autonomous, multi-step task execution and a larger connector ecosystem.

Key Capabilities

• Conversational chat with document upload (PDF, images, spreadsheets) and multi-turn context
• Web search grounding for up-to-date answers with citations
• Image generation via Black Forest Labs' Flux Pro model
• Voice interaction using Mistral's own Voxtral speech model for low-latency speech recognition and text-to-speech
• "Memories" — the assistant can recall prior conversations and stated user preferences to personalize future responses
• Agentic task execution (Vibe Work) that plans a task, asks for user sign-off, and then acts across connected tools (email, calendar, docs, tickets)
• Agentic coding (Vibe Code) that can scaffold multi-file features, write and run tests, and deploy changes from a CLI, VS Code, JetBrains, or Zed
• Connectors/MCP integrations to 60+ third-party services across categories like productivity, development, and commerce
• Custom MCP (Model Context Protocol) connector support for organizations that want to add their own tools

Main Features

Use Cases

• General-purpose question answering and research with web search grounding
• Summarizing, comparing, and extracting information from uploaded documents
• Drafting and editing written content (emails, reports, code comments)
• Generating illustrative or marketing images from text prompts
• Delegating multi-step office tasks (meeting prep, inbox triage, ticket updates) via connected apps
• Agentic software development: scaffolding features, running tests, and shipping small changes
• Enterprise deployments needing a European, GDPR-native alternative to US-based assistants

Supported Models / Technologies

As of September 2026, Vibe's default lineup draws on Mistral's own model family: Mistral Medium 3.5 (a frontier-class multimodal model tuned for agentic and coding use, with adjustable reasoning effort) and Mistral Small 4 (a hybrid model unifying instruction-following, reasoning, and coding, with a 256K token context window, merging the earlier Magistral, Pixtral, and Devstral lines). Document understanding uses Mistral OCR, voice features use Voxtral, and image generation uses Flux Pro from Black Forest Labs (a third-party partner model, not a Mistral model). Mistral updates these defaults periodically; the exact model version routed to a given plan is not always disclosed in the UI.

Platforms

• Web (chat.mistral.ai)
• iOS app
• Android app
• CLI (for Vibe Code)
• IDE integrations: VS Code, JetBrains, Zed
• API (La Plateforme, Mistral's separate developer API)

Integrations

Vibe's connector directory spans 60+ third-party services across productivity, development, and commerce categories, including GitHub, Slack, Notion, Linear, Sentry, Atlassian, Asana, Outlook and Outlook Calendar, Google Calendar, Gmail, Box, SharePoint, Databricks, Snowflake, Stripe, PayPal, Plaid, Square, and Zapier. Organizations can also register their own custom MCP (Model Context Protocol) connectors to ground the assistant on internal tools and data.

Pricing

Source: mistral.ai/pricing, cross-referenced September 26, 2026.

Advantages

• Built and hosted by a European company, which appeals to organizations prioritizing GDPR-native data handling
• Combines chat, agentic task execution, and coding assistance in a single subscription rather than separate products
• Broad, growing connector ecosystem (60+ integrations) plus support for custom MCP connectors
• Uses Mistral's own model family, which is independently benchmarked and documented
• Free tier includes image generation and a meaningful monthly API credit allowance

Limitations

• The 2026 rebrand to "Vibe" has caused confusion, since search demand and prior documentation still reference "Le Chat"
• Image generation depends on a third-party model (Flux Pro) rather than a Mistral-native image model
• Free tier limits on messages, web search, and coding sessions are restrictive for daily heavy use
• Team plan enforces a $50/month minimum regardless of seat count
• Enterprise pricing is not published and requires a sales conversation

Security & Privacy

Mistral is headquartered in the EU and states it operates under GDPR natively. On the Free tier, conversations may be used to improve models unless the user opts out in account settings; on Pro, Team, and Enterprise tiers, conversations are not used for training by default. Data submitted through the API is not used for training, and data accessed via connectors is fetched on demand rather than stored permanently. Enterprise customers can access a Data Processing Addendum, admin-managed training opt-in/opt-out controls, SAML SSO, and audit logs. For binding terms, consult Mistral's own privacy documentation and Data Processing Addendum rather than this summary.

Best For

Mistral Le Chat (Vibe) is best suited to individuals and organizations that want a single subscription covering general chat, document analysis, agentic office-task automation, and AI-assisted coding, especially those who value a European, GDPR-native vendor as an alternative to US-based assistants.

Alternatives

• ChatGPT (OpenAI)
• Google Gemini
• Anthropic Claude
• Microsoft Copilot
• Perplexity AI
• GitHub Copilot (for the coding-agent use case specifically)

Frequently Asked Questions

Is Mistral Le Chat still called Le Chat?

No. In 2026, Mistral AI rebranded Le Chat as "Vibe." Existing accounts, plans, conversations, and saved content carried over automatically, and the product is still reachable at chat.mistral.ai. Many people still search for and refer to it as "Le Chat," so this guide uses both names.

Is Mistral Le Chat / Vibe free to use?

Yes, there is a free tier with limited daily messages, limited web searches, limited agentic/coding sessions, image generation, and about $10/month in API credits. Paid Pro, Team, and Enterprise plans remove or expand these limits.

What AI models power Le Chat / Vibe?

As of September 2026, Vibe runs on Mistral's own model family, primarily Mistral Medium 3.5 and Mistral Small 4 for text and reasoning, Mistral OCR for document parsing, Voxtral for voice/speech, and Black Forest Labs' Flux Pro for image generation. Mistral periodically swaps in newer model versions, so the exact default model can change.

Can Le Chat / Vibe write and run code?

Yes. Its "Vibe Code" mode is an agentic coding assistant available via a command-line tool, a VS Code extension, JetBrains and Zed integrations, and a web interface, capable of multi-file feature scaffolding, running tests, and deploying changes.

Does Mistral train its models on my conversations?

On the free tier, conversations may be used to improve models unless you opt out in account settings. On Pro, Team, and Enterprise tiers, conversations are not used for training by default, and data sent through the API is not used for training either. Always confirm current terms in Mistral's own privacy documentation.

Final Summary

Mistral Le Chat — rebranded "Vibe" in 2026 — is Mistral AI's chat, productivity-agent, and coding-assistant product, built on the company's own Medium and Small model lines plus partner models for image and voice. It offers a free tier, three paid consumer/team tiers, and a custom enterprise tier, with a growing connector ecosystem and GDPR-native data handling as its main differentiators from US-based competitors.