Mistral Le Chat Review

Mistral built its reputation on open-weight models and a European address, and Le Chat was the consumer face of that pitch. In August 2026 the company folded the chat app and its coding agent into one product called Vibe, kept the same URL and login, and left the prices alone.

For a buyer the question is whether a sub-$15 plan from a European lab is a bargain or a compromise.

This review is written for professionals and small teams who already pay for ChatGPT, Claude or Gemini and want to know what they would gain or lose by switching, and for privacy-conscious users who want an assistant that does not train on their chats by default.

It covers what Vibe is, how the three modes work, every price Mistral publishes, who should buy Pro, who should stay free, and four alternatives with prices read on their own pages.

For the field as a whole, see our best AI chatbots for privacy ranking, where Le Chat appears, and our Claude review, ChatGPT review and Gemini review for the rivals.

Mistral Vibe, formerly Le Chat, is the AI assistant from Mistral AI, the European lab founded in 2023 that positions itself as Europe's sovereign alternative to the US model makers.

Mistral's help center describes the rename plainly: Le Chat became Vibe on August 12, 2026, with accounts, plans and conversations carried over, and the product now has three modes. Vibe Work is the productivity mode for delegating multi-step tasks across your apps.

Vibe Code is a coding agent available as a command-line tool, a VS Code extension and a web interface. Vibe Chat is the turn-based conversation Le Chat was known for.

Under the hood, Mistral's pricing page recommends Mistral Medium for most tasks and coding, OCR for documents, Voxtral for audio and speech, and Mistral Small for cost-sensitive projects. Voice input in Vibe runs on Voxtral.

The product page lists 100+ connectors (email, calendar, Slack, GitHub, Jira) with full MCP compatibility, deep research, spreadsheets and slides, scheduled tasks, image generation and a canvas for longer work.

What separates it from a ChatGPT clone is the ownership model: the models are Mistral's own, many are open-weight, enterprises can deploy on-premises or on Mistral Cloud with data residency, and the pricing table marks model training as opt-out on Free, Pro and Team alike. That combination is the reason to read this review rather than the price alone.

You open chat.mistral.ai or the mobile app and pick a mode. In Vibe Chat you ask a question and get an answer with optional web search and voice. In Vibe Work you hand over a task that spans tools, for example a one-page meeting brief built from your calendar, recent news and your knowledge sources, and Vibe runs the steps and reports back.

Scheduled tasks are capped at 5 on Free and unlimited on Pro and Team, which matters if you want recurring reports.

Vibe Code is where the product grew most. You install the CLI, launch it inside VS Code, JetBrains or Zed, or start a coding task in the web app and teleport it back to the terminal. Remote coding agents can write, test and open pull requests while you are offline, using context from GitHub, GitLab and Jira.

Free accounts get limited coding sessions; Pro unlocks what Mistral calls all-day coding in the CLI, IDE or web.

The limits are expressed as multiples rather than counts. Pro and Team give up to 6x the messages of Free, 5x the web searches and 40x the image generations, with 15GB of document storage on Pro and 30GB per user on Team. The only hard daily figure on the page is Flash answers: 150 a day on Pro and 200 on Team.

If you need to know exactly how many messages you get before you pay, Mistral will not tell you, and that is the first thing to test on the Free plan.

Mistral prices Vibe in US dollars on its own pricing page, read on October 17, 2026. Free is $0 with access on web and mobile, limited messages, web searches and coding sessions, basic image generation and 100+ connectors.

Pro is $14.99 a month, marked Popular, with more messages and web searches, faster responses, all-day coding, spreadsheets and slides, a monthly API credit allowance and chat and email support; verified students pay $5.99 a month for the same plan.

Team is $24.99 per user a month with even higher limits, 30GB of storage per user, domain verification and data export, and the plan page's calculator starts at $50 a month for the smallest team. Enterprise is a custom quote with private deployments, custom models and agents, audit logs, SAML SSO and white label.

All prices exclude taxes, per Mistral's pricing page; the Le Chat is now Vibe article confirms that plans carried over unchanged.

What a year costs. One person on Pro pays $179.88 a year on monthly billing, which is $60.12 less than a year of ChatGPT Plus or Claude Pro on monthly billing. A student pays $71.88 a year. A five-person team on Team seats pays $124.95 a month, or $1,499.40 a year.

The pricing page shows no annual discount for individuals, so the monthly figure is the figure. Developers pay separately through the API, where Mistral Large is listed at $0.50 per million input tokens and $1.50 per million output, with batch processing at half price.

Against the alternatives. ChatGPT's Plus plan is $20 a month with Go at $8 and three Pro tiers above it, per OpenAI's pricing page.

Claude Pro is $17 a month on an annual subscription ($200 billed up front) or the ChatGPT Plus price monthly, with Max from $100, per Claude's pricing page. Google AI Pro is $19.99 a month, AI Plus $4.99 and AI Ultra from $99.99, per Google's subscription page.

DeepSeek publishes no consumer plan price; its API lists deepseek-flash at $0.15 per million input tokens on a cache miss off-peak and $0.60 per million output, per DeepSeek's API pricing.

Mistral Pro is the cheapest full-featured paid tier in this group by $5 a month, and the student price is the lowest paid plan of any mainstream assistant here.

Toolradar data: on Toolradar, the software directory we run, Mistral AI is listed as an alternative to 15 other tools, ranking #2 (joint) of 36 in NLP Tools and #6 (joint) of 120 in AI Research, as of October 2026.

It sits at #124 (joint) of 1,557 AI Assistants on the same measure, far behind Claude at #1, which tells you Mistral is cross-shopped by developers and researchers more than by general chat users.

Vibe is a strong fit for professionals in Europe, privacy-conscious users anywhere, and developers who want a chat assistant and a coding agent in one subscription under $15. If you spend most of your day in VS Code or a terminal and want remote agents opening pull requests, Pro is priced below any rival that bundles the same.

Students get the same plan for $5.99, which makes it the obvious pick for anyone with a .edu address. Small teams that need SSO-free collaboration, domain verification and data export get a fair deal at the per-seat Team price.

It is a weaker fit if you need the frontier on hard reasoning, image or video generation, or the biggest plugin ecosystem. Mistral's own page positions Mistral Medium as the model for most tasks, and the feature table lists image generation but no video.

Usage limits are published as multiples of Free rather than counts, so heavy users cannot size a plan in advance. And if your workflow lives in Google Workspace or depends on custom GPTs, Gemini or ChatGPT will slot in with less friction. Skip it too if you want annual billing: the page shows none for individuals.