- → Mistral AI renamed Le Chat to “Vibe” on May 28, 2026 and repositioned it as a complete work tool
- → Work Mode connects to Google Workspace, Outlook, SharePoint, Slack and GitHub to run tasks independently
- → Code Mode launches code agents in parallel in isolated cloud sandboxes, until the pull request
- → 4 prices: Free, Pro €14.99, Team €24.99/user, Enterprise on quote. Students -50% on Pro
Mistral announces: ASIC, Airbus, BMW and Vibe (via Fabienne Billat on X)
Mistral AI is undergoing its biggest change since the launch of Chat. The chatbot has a new name (Vibe), a new identity (work tool), and two modes that put it in direct competition with ChatGPT Agents, Claude and Gemini. Above all, this is the first time that a European player has offered a complete agentic platform on a large scale. We’ll break down what it actually changes, and who it’s really for.
01 · The rebrand
Why Mistral renames Le Chat to Vibe
Officially, it is an accepted marketing decision: “Le Chat” sounded too chatbotin a world where real uses drift towards the agentic. Vibe, the new name, indicates the positioning: an assistant that vibes with your working day, not a conversational agent that you consult occasionally.
Unofficially, there is another reason. The name “Vibe” already existed at Mistral, attributed to their coding tool launched last year (Vibe 2.0, with paid Pro and Team access). By reusing the name for the entire product, Mistral consolidates its range: a single word, a single product, several modes. Where OpenAI multiplies names (ChatGPT, GPT, Codex, Operator, etc.), Mistral simplifies.
Existing conversations, settings, and memory are all automatically transferred to Vibe. No migration, no new app to install. The rebrand is cosmetic for the basic user, structural for the strategy.
02 · Work Mode
Work Mode: Vibe connects to your professional tools
This is the bravery piece of the launch. Work Mode allows Vibe to connect to your work environment and perform end-to-end tasks, not just answer questions.
Google Workspace Gmail, Drive, Docs, Sheets. Vibe can scan your mailbox, produce a summary, modify a table, send an email.
Microsoft Outlook + SharePoint Identical on the Microsoft side. Scanned Outlook mailboxes, opened SharePoint documents, modified files.
Slack Retrieving channels, searching for discussions, posting messages. Ideal for summarizing a project conversation.
GitHub Reading issues, opening PR, tracking commits. Preparing the ground for Code Mode.
Before starting a task, Vibe presents you his action plan and awaits your approval. Each step is then expandable and verifiable. Tasks can be recurring (daily, weekly, monthly), which makes it a mini-Zapier controlled in natural language. And reusable workflows can be saved as “skills,” a nod to the Skills launched by Anthropic last year.
The European ETI angle is clearly at play. Mistral is aimed at companies that want a AI agentAn autonomous system that perceives its environment, makes decisions and executes actions to achieve a goal. able to work in their Microsoft 365 or Google Workspace stack, hosted in Europe, compliant GDPRGeneral Data Protection Regulation, the European framework on digital privacy.. This is the main argument against ChatGPT Enterprise.
03 · Mode Code
Code Mode: agents code in parallel, up to the pull request
The other new feature of Vibe is its developer mode. Code agents rotate in isolated cloud sandboxes. Everyone can develop a feature, fix a bug, write tests and open a pull request, without blocking others.
The benefit of parallelism is immediate: you launch 5 tasks simultaneously, you close your laptop, and you retrieve the pull requests the next day. The sessions survive the closing of the laptopwhich is one of the main frustrations of Cursor Agents and Claude Code in local mode. From June, Mistral also promises the launch of Code Mode sessions directly from Slack.
Mistral also delivers a new VS Code extension and an update of its CLI with a command /teleport which moves current sessions (and their history) between the local terminal and the cloud. This is the detail that proves that Mistral knows its dev audience: the terminal-cloud transition without loss of contextThe context window: the maximum number of tokens that a model can process in a single request (eg: 200k, 1M). is the pain point n°1 of coding agents 2026.
04 · Prices
The 4 Vibe prices: the price of agent autonomy
Vibe arrives with four levels, the details of the usage limits of which remain partially opaque (Mistral communicates in multiples of the Free plan rather than in absolute figures).
Free
0 €/month
Limited use, basic access to Work Mode and Code Mode
Pro
€14.99/month
×6 messages, ×40 images, 15 GB of storage, students at €7.49
Team
€24.99/user/month
Same quotas as Pro + admin, data export, shared spaces (€19.99 annually)
Enterprise
On quote
SLA, dedicated support, sovereign hosting, advanced integrations
The interesting detail: Pro and Team have the same usage quotas. The €10 extra per user on Team does not pay for additional volume, but for admin tools (team management, shared spaces, domain verification, export). This is consistent with the target: Team is designed for SMEs, Enterprise for large accounts with sovereignty needs.
05 · Competition
Vibe versus ChatGPT Agents, Claude and Gemini
Mistral does not invent the wheel. The agent + coding sandbox + integrations configuration already exists elsewhere. OpenAI has Codex and Operator. Anthropic has Claude Code in the browser. Google has Gemini with Workspace built-in. Cursor even has an agent fleet mode in version 3.
The differentiating angle of Mistral can be summed up in two words: European and Bundle. Where ChatGPT charges Plus, Codex, and enterprise agents separately, Vibe delivers everything in one subscription. Where OpenAI and Anthropic host in the US by default, Mistral offers sovereign European hosting for the Team and Enterprise versions.
The model angle also plays a role. Vibe runs on Mistral Medium 3.5, which has a toggle reasoning_effort allowing you to switch between quick mode and more advanced mode. This is the same approach as ChatGPT (with its reasoning control) and Claude Opus 4.8 (with its Fast Mode). The ecosystem is converging towards this paradigm: one model, two speeds, the user chooses.
If you want to compare in detail, our comparison ChatGPT vs Claude vs Gemini vs Mistral covers the general public side. And for the complete Mistral technical sheet, the model page follows developments week by week.
06 · For whom?
Who is Vibe really for?
On paper, Vibe is aimed at everyone. In practice, three profiles really benefit from the new positioning.
The first is the user of a Microsoft 365 or Google Workspace environment who wants an AI agent capable of moving their files, emails and tasks without copying and pasting. This is where the use value is most immediate, and it is Mistral’s number one target for this campaign.
The second is the French or European developer who is looking for an alternative to Claude Code and Cursor with sovereign hosting. Code Mode is less mature than Claude Code 2.0 or Cursor 2, but it saves time at the same time and it integrates with Slack and VS Code.
The third is the French ETI or SME which seeks to equip its teams with AI without going through the American ecosystem. Mistral has already signed with La Banque Postale, Caisse des Dépôts, Airbus and several ministries. Vibe is the consumer product that consolidates this B2B strategy.
For individuals and pure freelancers, the Pro plan at €14.99 remains equivalent to ChatGPT Plus or Claude Pro, with a little less raw capacity on the reasoning side, but the sovereignty argument in addition. And the Free plan is a good entry point to test Work Mode without commitment.
07 · And after?
Mistral and the bet of the complete agent ecosystem
With Vibe, Mistral is taking on a clear challenge: become the European OpenAInot one model among others. The rebrand is a strategic signal. The company is no longer just a supplier of LLMLarge Language Model, a language model trained on billions of parameters to generate and understand text. open-weights is a complete product player, with a general public assistant, a worker agent, a coding agent and soon (June 2026) a native integration in Slack.
The timing is not trivial. Mistral has just raised 3 billion euros at a valuation of 22 billion, announced a 1.4 GW campus in Paris with the acquisition of Emmi AI and signed a series of strategic contracts. Vibe is the product showcase of all this rise in power.
The risk is to disperse. ChatGPT, Claude and Gemini each have their clear strength (reasoning, code, Google integrations). Vibe wants to do everything at once. Success will depend less on technology than on Mistral’s ability to penetrate the daily use of French SMEs and mid-sized companies. So far, it’s a great launch. The verdict will come in 6 to 12 months, as the uses take root or not.
To go further
To understand the complete Mistral ecosystem, the Mistral model page lists all current versions, prices and benchmarks. To evaluate Vibe against its competitors, the comparison of the 4 major LLMs is regularly updated. And for the business context of the fundraising which financed all of this, our file on the €22 billion valuation explains the strategic background.

