The short version. Mistral rebranded Le Chat to Vibe and turned it into an agent platform. Two modes: Work mode for knowledge work and admin (Gmail, Drive, Slack, Outlook, SharePoint, GitHub integrations), Code mode for development (web, terminal, IDE, sandboxed code execution). Free tier, $14.99 Pro, $24.99 Team, custom Enterprise.
The positioning. Mistral is bundling its full stack (Mistral Large 3, Mistral Small 4, Voxtral) under a single user-facing product. Direct competition to ChatGPT Plus, Claude Pro, and Gemini Advanced.
Vibe is Mistral’s bet that the next chat product is not a chat product. It is an agent that does the work, with the chat as the steering interface. ChatGPT, Claude, and Gemini are converging on the same bet from the other direction.
Mistral retired the Le Chat brand and replaced it with Vibe. The new product is positioned as an agent platform rather than a chat assistant. The agent operates in two modes (Work and Code), integrates with the standard enterprise productivity stack, and ships at consumer-friendly pricing. This guide breaks down what Vibe actually does, how it compares to the existing agent products, and whether non-technical readers should consider switching from their current AI tool.
What is Mistral Vibe?
An AI agent platform from Mistral AI. Launched at chat.mistral.ai with iOS and Android mobile apps. Runs on Mistral’s flagship models (Mistral Large 3 and Mistral Small 4) underneath the agent layer.
Vibe replaces Le Chat as Mistral’s primary consumer-facing product. Existing Le Chat accounts migrate automatically. The product was unified to consolidate Mistral’s various consumer experiences into one place with one billing plan and one feature roadmap.
What are Work Mode and Code Mode?
Work Mode. The agent handles multi-step business work: enterprise knowledge search, structured data analysis, document synthesis, scheduled task automation. Integrates with Google Workspace, Outlook, SharePoint, Slack, GitHub. Permissions are scoped by admin policies. Use case: “summarize last quarter’s sales pipeline from HubSpot, draft a board update, and put it on my calendar to review Tuesday morning.”
Code Mode. The agent handles development workflows: feature building, bug fixes, refactoring, pull request generation. Runs in web, terminal, and IDE environments. Sandboxed code execution requires user approval for sensitive actions. Sessions persist across system restarts, which means you can step away and come back. Direct competition to Claude Code and OpenAI Codex.
The split is intentional. Most consumer agent products today force users into a single conversation thread. Vibe separates concerns so a marketing manager can stay in Work Mode and never see code, while a developer can stay in Code Mode without admin productivity clutter.
How does Vibe compare to Claude and ChatGPT?
- vs. ChatGPT Plus ($20/month). Vibe is $5 cheaper ($14.99 Pro). ChatGPT has the plugin/GPT ecosystem and longer-running consumer brand. Vibe has stronger European data residency story. Quality of underlying model: ChatGPT GPT-5 still leads on frontier benchmarks.
- vs. Claude Pro ($20/month). Vibe is cheaper. Claude has the best-in-class long-context handling and the Claude Code product is more mature. Vibe matches Claude on the agent-platform positioning but is newer to it.
- vs. Gemini Advanced ($20/month). Vibe is cheaper. Gemini has the Google Workspace integration depth Vibe is just building. Vibe is platform-agnostic.
- vs. Mistral Vibe Team ($24.99/user/month). Bundles shared workspace, admin controls, and Mistral’s open-weight option for enterprises that need self-hosting.
For a fuller comparison of the consumer AI chat tools, see ChatGPT vs Claude vs Gemini.
What can Vibe do that other agents cannot?
Three differentiators worth knowing.
- Session mobility. Move a live session between terminal and cloud without losing context. Useful for developers who start a Code Mode session locally and want to continue from a phone or a different machine.
- Reusable skills via open standards. Vibe supports the same open agent skill formats that Claude and others are converging on (MCP-style protocols). This means a skill written for one agent should be portable to another. Less vendor lock-in than the early agent products had.
- Scheduled task automation. Vibe can run agents on a schedule, not just on demand. “Every Monday morning, summarize Friday’s email and post to my Slack.” This is the recurring-workflow feature most consumer agent products have not yet shipped at scale.
Should I switch from ChatGPT or Claude to Vibe?
Probably not yet, with two exceptions.
The first exception: if you are an EU-based individual or organization with data residency requirements, Vibe is the natural choice. Mistral is the only major AI lab headquartered in Europe; data processing within EU jurisdictions is the strongest selling point.
The second exception: if cost matters more than frontier-model quality. Vibe Pro at $14.99 vs. ChatGPT Plus at $20 and Claude Pro at $20 is real money over a year. The underlying Mistral models are mid-tier in quality; for most everyday tasks the difference is minor.
For everyone else, the recommendation in 2026 is to stay with the tool you know. Switching costs (learning new prompts, rebuilding workflows, retraining muscle memory) usually outweigh the marginal product differences. See How to use Claude AI if you are still picking your first tool.
Where does this leave the Mistral product family?
Vibe is the user-facing layer on top of Mistral’s full model stack: Mistral 3 family for the flagship, Mistral Small 4 for the unified mid-tier, Voxtral for voice. Vibe pulls all of those into a single chat and agent surface.
The pattern is the same one OpenAI, Anthropic, and Google have settled on: own the model, own the consumer product, own the developer API. Mistral now has all three in production. Whether they can win significant share against the US labs is the question 2027 will answer.
Frequently asked questions
What happened to Le Chat?
Rebranded to Vibe. Existing accounts migrated automatically. The Le Chat URL forwards to chat.mistral.ai. No data loss, no separate signup.
Is Vibe free to try?
Yes. The Free tier covers basic tasks and simple queries. The Pro tier ($14.99/month) adds complex reasoning and full-day coding. Team adds shared workspace and admin controls.
How is Vibe Code Mode different from Claude Code?
Claude Code is terminal-first and CLI-driven. Vibe Code Mode runs in web, terminal, and IDE environments with persistent sessions across all three. Both products do agent-style coding work with sandboxed execution. Claude Code has the more mature ecosystem of plugins and skills; Vibe is newer to the category.
Can Vibe replace my CRM or my project management tool?
No. Vibe Work Mode reads from those tools and acts on top of them, but does not replace them. Your CRM stays your system of record; Vibe is the agent that operates on top of it.
Is Vibe safe for enterprise use?
Mistral offers Team and Enterprise tiers with admin controls and audit logs. Data residency stays within EU jurisdictions by default. For agent-safety considerations broadly, see Zero Trust for AI Agents.
Get the daily Beginners in AI newsletter
One issue a day. Plain English coverage of agent products and the trade-offs that matter when you are picking your stack.
Get Smarter About AI Every Morning
Free daily newsletter. Built for people who want to use AI well, not chase every model.
Free forever. Unsubscribe anytime.
Sources
- Mistral: Vibe agent platform announcement
- Mistral Vibe consumer app
- Mistral API documentation
- Model Context Protocol (MCP) standard
You might also like
- Mistral Small 4 Explained
- Mistral 3 Family Explained
- Mistral Voxtral TTS
- ChatGPT vs Claude vs Gemini
- Zero Trust for AI Agents
- AI Glossary
Two ways to go further
The AI Prompt Library
1,000+ ready-to-use prompts for Claude, ChatGPT, and Gemini. Stop staring at a blank box.
Get it for $39 →2-Hour Live AI Crash Course
A private, beginner-friendly session across Claude, ChatGPT, Gemini, and the wider landscape.
Book for $125 →