Level : Beginner
Mistral Le Chat: Europe's AI Assistant Takes On the Giants
By The Info Agent IA editorial team ·
Le Chat, the assistant from French startup Mistral AI, is now called Vibe, and it remains the leading European alternative to ChatGPT, Claude and Gemini. It handles nearly every everyday task for less money, but its models still trail the very best, and its default settings deserve a closer look.
Le Chat, now Vibe: what is it?
Le Chat is the chat assistant made by Mistral AI, an artificial intelligence lab based in Paris. It launched in February 2024 and was renamed Vibe (or Mistral Vibe) in 2026. In practice, nothing changes for you: the address is still chat.mistral.ai, and accounts and conversations carried over. We call it Vibe below, but plenty of people still say Le Chat.
Vibe is now split into three spaces:
- Chat, the classic conversation: ask a question, write, summarize, translate.
- Work, for your job: the assistant plugs into your apps through connectors.
- Code, for developers, with a command-line tool and an extension for the VS Code editor.
It runs on the web, on iPhone and on Android. Mistral AI is more than an app maker: it builds its own language models and releases some of them as open weights, meaning you can download them and run them on your own servers.
What Vibe can do, next to the giants
Here is what each company lists on its official pages as of October 22, 2026, for Vibe and its three big American rivals:
| Assistant | Free | Paid from | Web search | Agent mode | Voice | Images |
|---|---|---|---|---|---|---|
| Le Chat (Mistral Vibe) | Limited | $5.99/mo | Yes | Yes | Yes | Yes |
| ChatGPT | Limited | $8/mo | Yes | Yes | Yes | Yes |
| Claude | Limited | $20/mo | Yes | Yes | — | — |
| Gemini | Limited | €4.99/mo | Yes | — | Yes | Yes |
On paper, Vibe holds its own. Mistral AI lists web search, an agent mode that can chain several steps, file uploads, image generation, voice, memory across chats and code execution. That is one of the most complete feature lists on the market.
A concrete example: you are preparing a community group meeting. You upload the minutes of the last one, ask for a list of open items, then for an agenda and an invitation email. That kind of task does not need the most powerful model in the world, and Vibe handles it with ease.
The Work space goes a step further. Once your apps are connected, the assistant can fetch information where it lives instead of you copying it by hand. Before you connect your inbox or a shared folder, though, ask yourself whether you are comfortable with an assistant reading it. It is the same question you would face with Gemini in Gmail or Copilot in Word.
Plans and pricing
According to Mistral AI’s pricing page, checked on October 22, 2026, these are the available plans. Prices are shown in US dollars before tax; the site also offers a display in euros.
| Plan | Price | Best for |
|---|---|---|
| Free | Free | Limited messages, searches and images |
| Student | $5.99 a month | Verified students, up to 12 months |
| Pro | $14.99 a month | Regular use, with $15 of API credits a month |
| Team | $24.99 per user per month | Teams, $50 a month minimum |
| Enterprise | Custom quote | Large organizations |
For comparison, in the same survey ChatGPT Plus and Claude Pro cost $20 a month. Vibe Pro is therefore about a quarter cheaper, and it includes credits to use Mistral’s models in your own tools.
The Student plan stands out: at $5.99, it is one of the most affordable general-purpose AI subscriptions around. Our guide to AI for students puts it alongside the other options. If you are not sure you want to pay at all, our roundup of the best free AI compares what each assistant gives you without a subscription, and our analysis of which AI subscription is worth the money can help you decide.
Mistral’s models: good, but not on top
Mistral AI does not state on its product page which model powers Vibe at any given moment. We can, however, look at its flagship models, available to developers:
| Model | Developer | Release | Context (tokens) | Input ($/M tokens) | Output ($/M tokens) | Open weights |
|---|---|---|---|---|---|---|
| Mistral Large 4 | Mistral AI | October 6, 2026 | 1.05 M | $1.36 | $4.18 | No |
| Mistral Medium 3.5 | Mistral AI | April 30, 2026 | 262 k | $1.5 | $7.5 | No |
| Mistral Small 4 | Mistral AI | March 16, 2026 | 262 k | $0.16 | $0.63 | Yes |
| Devstral 2 | Mistral AI | December 9, 2025 | 128 k | $0.40 | $2 | Yes |
The newest, Mistral Large 4, came out on October 6, 2026, with a very large context window (over a million tokens) and a very low price for a model of its class. Here is how it compares with the two top models of the moment:
- Claude Sonnet 5.595.9
- Gemini 4 Argon95.3
- Mistral Large 479.7
The gap is real on hard tasks: long reasoning, advanced math, complex programming. When you are drafting an email, summarizing a document or looking for a dinner idea, you will notice it much less. Another plus: several Mistral models, including Mistral Small 4 and Devstral 2, are released as open weights. You will find them in other services too, such as DuckDuckGo’s Duck.ai or the Leo assistant in the Brave browser.
The European edge: hosting and legal framework
This is the argument you hear most often. Mistral AI says it hosts data in the European Union by default. Your conversations therefore fall under the GDPR, Europe’s data protection law, and stay on servers located in Europe. For a business, a public body or a professional bound by confidentiality, that is often the deciding factor.
There is a bigger picture too. Investment figures show how unevenly resources are spread between Europe and the United States:
| Private investment in AI | 2024 | 2025 |
|---|---|---|
| United States | $94.6 billion | $240.6 billion |
| Europe | $16.9 billion | $17.6 billion |
| China | $8.1 billion | $10.4 billion |
Source: Our World in Data, private investment in artificial intelligence series, in US dollars.
In our view, this table goes a long way toward explaining why Mistral AI still trails the best American models: Europe invests roughly fourteen times less. Picking Vibe is also a way to back a European player, a fair motivation that our piece on European AI explores further.
If privacy is your top priority, other European options exist. Lumo, from Proton, runs on servers in Europe and does not train any model on your conversations, but it offers fewer features.
Who is Vibe for?
| Your profile | Is Vibe a good pick? | Why |
|---|---|---|
| You want an AI hosted in Europe | Yes | EU data hosting by default |
| You are a student | Yes | Student plan at $5.99 a month |
| Tight budget, but you want a subscription | Yes | Pro is cheaper than ChatGPT Plus or Claude Pro |
| You build with open models | Yes | API credits included, open-weight models |
| You want the most capable model | No | Mistral Large 4 trails Claude and Gemini |
| You live in Google or Microsoft 365 | Probably not | Gemini or Copilot integrate better |
Check yourself You run a small accounting firm, clients keep asking where their data is stored, and you mostly use AI to draft letters. Does Vibe make sense?
Yes, broadly. EU hosting by default answers your clients’ question, and drafting letters does not require the most powerful model. Do remember to turn off training on your conversations, or move to the Team plan, and never paste sensitive client data without checking the terms first.
Limits to keep in mind
- Models that trail the leaders. On the hardest tasks, Claude, Gemini and ChatGPT keep a measurable lead.
- An unnamed model. Mistral AI does not clearly say which model answers you in Vibe, which makes comparisons harder.
- A default setting that is not very protective. Training stays on until you turn it off.
- A smaller ecosystem. There is no in-house office suite as at Google or Microsoft, although the Work space connectors partly make up for it.
- A name in flux. The switch from Le Chat to Vibe can be confusing when you look for help or tutorials.
And like any assistant, Vibe can be confidently wrong: always double-check important figures, dates and references.
For a point-by-point face-off, see our ChatGPT vs Le Chat comparison. To see where Vibe fits among all assistants, read our analysis of which AI is the best, or compare features yourself with the AI assistants comparison tool.
Frequently asked questions
Does Mistral's Le Chat still exist?
Yes, under a new name. In 2026, Mistral AI renamed Le Chat to Vibe (Mistral Vibe). The address is still chat.mistral.ai, and existing accounts and conversations were kept.
How much does Mistral Vibe cost?
According to Mistral AI's pricing page as of October 22, 2026, there is a limited free tier. Pro costs $14.99 a month before tax, Student costs $5.99 a month for up to 12 months, and Team costs $24.99 per user per month.
Is my data stored in Europe?
Mistral AI says it hosts data in the European Union by default. Be aware, though, that training on your conversations is turned on by default for the Free, Pro and Student plans. You can switch it off in the privacy settings.
Is Vibe as good as ChatGPT or Claude?
For everyday tasks, it covers the essentials. But on Epoch AI's index as of October 22, 2026, Mistral Large 4 scores 79.7 out of 100, versus 95.9 for Claude Sonnet 5.5 and 95.3 for Gemini 4 Argon. On the hardest tasks, the gap is clear.