Understand Mistral Le Chat's move to Vibe, current plans, connected workflows, a published hands-on experience, and the privacy controls to check.
Mistral Le Chat is now Vibe. If you search for Le Chat or open its familiar address, you are looking for Mistral's assistant under a newer name. The service has grown from a conversational interface into a workspace for research, documents, connected tasks, and coding.
The important question is what you can finish with it. A quick explanation, a sourced briefing, and an agent working across business tools need different instructions and different levels of review.
This guide explains the current product, pricing, privacy controls, and a practical workflow. It keeps the Le Chat name where it helps readers find the product, while describing the latest documented experience.
Documentation and USD prices checked October 9, 2026. Belliste has not conducted an original hands-on benchmark. The independent experience below is attributed to its author and date.
Le Chat was Mistral AI's conversational assistant, introduced in February 2024. Mistral's original announcement now directs readers to Vibe.
The company offers its own model families, a hosted assistant, developer APIs, and enterprise deployment options. These are related products, but a chat subscription is not the same thing as an API contract or a self-hosted model license.
You can still start at chat.mistral.ai. The product's name changed; the familiar entry point remains.
Mistral's rename guidance says existing accounts, plans, conversations, and saved content carry over.
The newer Vibe overview documents another change: Chat and Work began merging on September 22. The unified web and mobile workspace handles both quick conversation and more involved tasks, using a fast/think control. Code remains a development experience.
Migration is progressive, so an older screenshot or account may still show separate surfaces. Enterprise administrators have a separate migration window.
| Name you may encounter | What it means |
|---|---|
| Le Chat |
| Vibe Work | The unified web and mobile workspace |
| Fast / think | Controls for task depth in the current workspace |
| Vibe Code | Coding through supported terminal, editor, and remote interfaces |
| Mistral Studio | Developer tooling and API access |
Do not interpret an old tutorial's interface as proof that a feature has disappeared.
Starting price: Free $0/month · Pro from $14.99/month.
Mistral's current pricing page lists:
| Plan | Listed USD price | Main purpose |
|---|---|---|
| Free | $0 | Try the assistant and limited tool usage |
| Pro | $14.99/month | More usage and deeper individual work |
| Team | $24.99/user/month | Collaboration and organizational controls |
| Enterprise | Custom quote | Private deployments and tailored requirements |
Paid prices are listed excluding taxes and subject to fair usage. Education offers have eligibility conditions; verify them at checkout.
The comparison lists up to 15 GB of document storage for Pro and 30 GB per user for Team. Relative allowances are described as “up to” multiples of Free, rather than a universal unlimited quota.
Check the live subscription screen before budgeting. A paid plan can still have feature limits, and usage beyond included allowances may involve pay-as-you-go charges.
Mistral's Vibe announcement describes research, document drafting, and multi-step work across connected tools, alongside dedicated coding workflows.
Useful starting points include:
An assistant can produce an answer or a deliverable without proving either is correct. Define what successful work looks like before starting.
In his July 10, 2026 TechRadar review, Ritoban Mukherjee described testing the Pro plan over several weeks.
He reported that Work Mode efficiently handled multi-step research using connected sources and produced a structured document. In a coding task, it scaffolded an API integration and prepared a draft pull request, with him stepping in twice to provide direction.
He also criticized unclear usage limits and limited onboarding. Those observations make the review more useful than a simple list of features: they describe where the workflow needed human help.
This was a dated test before the latest documented interface merger. It is not a benchmark of every current model or account.
Our editorial takeaway is to test a complete task. Count the corrections and verification work, not just the time to the first answer.
Start with a source you understand: a report, a meeting note, or a spreadsheet. Ask for a bounded deliverable, such as a one-page briefing or a list of inconsistencies.
Before uploading client or internal documents, check the training controls and your organization's policy. A European provider does not automatically make every hosted conversation private or suitable for every dataset.
Attach the necessary material and identify which file matters. State the audience, deadline, scope, and facts that must remain unchanged.
A straightforward explanation may need only a quick response. A research brief needs source checking and a more deliberate process. The current workspace can look different during migration, so follow the controls your account actually shows.
Connect only the services the task needs. Ask for a draft before any external action, and review the proposed recipient, destination, and content.
Open original pages and compare summaries against the source. Check numbers, quotations, dates, and whether a recommendation follows from the evidence.
Give specific feedback, then inspect the saved document or spreadsheet. A confident chat response does not guarantee correct formatting, formulas, or references.
Outcome + Sources + Audience + Constraints + Evidence rules + Format
> Using the attached supplier proposals, draft a one-page recommendation for a small purchasing team. Compare total cost, delivery dates, and support terms. Cite the relevant proposal sections, mark missing information, and do not invent prices or choose a winner without explaining the trade-offs.
For writing:
> Rewrite this customer email in clear English. Preserve the order number, delivery date, and refund terms. Give two versions under 120 words and flag any promise that is not supported by the original.
For research:
> Research this topic using official documentation first. Separate confirmed facts from inference, include publication dates, and list questions that remain unresolved. Return a brief I can verify in ten minutes.
For coding:
> Investigate this failing behavior, explain the cause, and propose the smallest change. Run the relevant checks and show the remaining limitations before treating the task as complete.
Web search is useful when your question depends on current information. Uploaded files are useful when the answer must reflect a particular source. Connected tools are useful when the relevant material is scattered.
These inputs have different failure modes. Search can return an outdated page. A document can omit a needed detail. A connector can expose more information than the immediate task requires.
Ask the assistant to identify its source and explain what it cannot establish. For recurring work, keep a stable instruction set and a known example of a correct output. Reuse helps only when the underlying facts are checked again.
Mistral's training guidance says standard Vibe inputs and outputs are used for training by default unless you opt out. Enterprise customers are opted out by default, with an administrator-controlled opt-in.
The opt-out instructions identify uploaded documents as input data. In the Admin panel, open Vibe's Privacy section and disable the training toggle. Mobile apps have their own Data & Account Controls.
Vibe and API training switches are separate. Changing one does not automatically change the other. The guidance also treats submitted feedback separately, so check what sharing a rating and comment authorizes.
The current overview describes Temporary Chat as excluded from conversation history and model training, with legal-retention exceptions.
That is different from API Zero Data Retention. Mistral's ZDR guidance limits that option to approved pay-as-you-go use of supported stateless API endpoints. It does not apply to ordinary Vibe Work or Chat, libraries, or other stateful products.
Training, saved history, and API retention answer different questions. Review each one instead of treating “not used for training” as a promise that no data is stored.
The hosted assistant and its underlying models should be evaluated separately. A model card's context window or benchmark result does not establish the limit or performance of every task in the consumer interface.
For developers, use Mistral's documentation to select the exact model and endpoint. Review token or tool billing, retention, and training controls for that route.
For local deployment, inspect the license attached to the specific weights. Do not assume that the hosted assistant can simply be downloaded with all its connectors, interfaces, and management features.
Consider it if you want a Mistral-powered assistant for research, writing, documents, and connected workflows, with a relatively accessible Pro subscription.
For teams, assess the administrative controls you actually need. For individuals, compare the free and paid experiences on the same recurring task.
A useful evaluation has three measures: factual accuracy, time spent correcting the output, and whether the permissions and data controls fit your work. Fast generation is valuable when the result survives those checks.
Le Chat's familiar name now leads to Vibe, a broader assistant for conversation, professional tasks, and coding. The latest documentation matters because the interface and product terminology have continued changing.
Start with a small source-based task, configure privacy settings, and inspect the final result. Upgrade when a recurring task benefits from the additional tools or allowance. Open Mistral's assistant to try it.
Le Chat has become Vibe. The chat.mistral.ai entry point remains, and Mistral says existing accounts, plans, conversations, and saved content carry over.
Mistral documents a progressive merger of Chat and Work starting September 22, 2026. Quick conversation and deeper tasks share a workspace with a fast/think control; Code remains a development experience.
A Free plan is available with limited usage. Paid plans offer more capacity and features, but fair-usage and feature limits still apply.
The USD pricing checked October 9, 2026 lists Pro at $14.99 monthly and Team at $24.99 per user monthly, excluding taxes. Enterprise uses custom pricing.
Mistral says standard Vibe input and output data are used by default unless you opt out. Enterprise customers are opted out by default. Uploaded documents count as input data.
No. Mistral documents separate Vibe and API training controls. Configure the setting for each service you use.
Mistral describes it as a conversation not saved to history or used for training, with possible legally required retention. It is distinct from API Zero Data Retention.
No. Mistral limits ZDR to approved pay-as-you-go use of supported stateless API endpoints. It does not apply to ordinary Vibe Work or Chat and other stateful products.
