Skip to content
trending

Guides

Mistral AI Pricing: Vibe Pro & API Costs in 2026

Mistral Vibe (formerly Le Chat) offers a free plan, while Pro costs $14.99/month. API pricing starts at $0.10 per million tokens. Here's everything you need to know.

By 4 min read
Mistral AI Pricing: Vibe Pro & API Costs in 2026

👁 117 reads

Mistral Vibe (formerly Le Chat) has a free plan, and Pro costs $14.99 per month on Mistral’s website. Mistral’s API starts at $0.10 per million tokens (Ministral 3 3B), while the flagship Mistral Medium 3.5 costs $1.50 for input and $7.50 for output. We’ve covered the models and chat interface in our Mistral AI guide, and here we focus solely on the money: how much, for what, and when you don’t need to pay at all.

As of October 3, 2026

  • Vibe Free: $0, with message limits and $10 API credits per month.
  • Vibe Pro: $14.99 per month (mistral.ai).
  • Student: $5.99 per month, maximum 12 months, after verification.
  • Team: $24.99 per user per month, $50 minimum.
  • Cheapest API model: Ministral 3 3B at $0.10 per million tokens for input and output.

Mistral Vibe (Le Chat) Pricing

Le Chat changed its name to Mistral Vibe in May 2026. Your account, conversations, and subscription carried over unchanged, so if you were paying for Le Chat Pro, you’re now paying for Vibe Pro. The pricing on Mistral’s website is in US dollars:

Plan Price on mistral.ai What you get
Free $0 limited messages, searches, and coding sessions, $10 API credits per month, over 100 connectors (email, calendar, drives)
Pro $14.99 per month more messages and searches, “all-day” coding in terminal, editor, and browser, $15 API credits per month, chat and email support
Student $5.99 per month same as Pro plus $30 API credits per month; requires student verification, maximum 12 months
Team $24.99 per user per month ($50 minimum) shared workspace, up to 30 GB per user, domain verification, data export
Enterprise custom pricing custom models and agents, SAML SSO, white label, audit logs, on-premises deployment

Mistral doesn’t specify exact Free plan limits (how many messages per day), only that Pro gives you more. From our testing, Free is sufficient for a dozen or so regular questions daily, but we didn’t measure this systematically, so treat that as an observation rather than a hard number.

Vibe Pro Compared to Competitors

Service Plan USD Price (October 3, 2026)
Mistral Vibe Pro $14.99 per month
ChatGPT Plus $20.00 per month
Claude Pro $20.00 per month

Vibe Pro is cheaper than ChatGPT Plus and Claude Pro. The reason to pay Mistral specifically is that it’s a European provider with data processed in Europe, rather than output quality. A full comparison of AI subscriptions is in our AI subscription pricing article.

Mistral API Pricing

If you’re building something custom or connecting Mistral to your own software, you pay per token. Prices per million tokens, in US dollars, excluding VAT:

Model Version Input (USD) Output (USD) Weight License
Mistral Medium 3.5 26.04 1.50 7.50 modified MIT
Mistral Large 3 25.12 0.50 1.50 Apache 2.0
Mistral Small 4 26.03 0.15 0.60 Apache 2.0
Ministral 3 14B 25.12 0.20 0.20 Apache 2.0
Ministral 3 8B 25.12 0.15 0.15 Apache 2.0
Ministral 3 3B 25.12 0.10 0.10 Apache 2.0
Codestral 25.08 0.30 0.90 proprietary
Devstral 2 25.12 0.40 2.00 not verified
Voxtral Small 25.07 0.10 0.30 Apache 2.0

We read these prices on October 3, 2026, from OpenRouter’s model list, where Mistral itself is the provider. On mistral.ai/pricing we confirmed the rate for Mistral Large 3 ($0.5 and $1.5) directly; the others can’t be read there without logging into the console. Batch processing is half price - for example, Medium 3.5 costs $0.75 and $3.75 in batch mode.

Interesting note from the table: Mistral Large 3 is three times cheaper than Mistral Medium 3.5 on input and five times cheaper on output. The “Large” name doesn’t mean the most expensive model here. Medium 3.5 is the newer model from April 2026 that replaced specialized code and reasoning models (Devstral and Magistral), and it’s now at the top of Mistral’s offering.

What It Costs in Practice

We ran the same five prompts in Polish (summary, email, correction, explanation for a child, translation) through four Mistral models via API. The bill from OpenRouter for all five tasks:

  • Mistral Medium 3.5: $0.0061 for 5 tasks (634 response tokens),
  • Mistral Large 3: $0.0015 (729 tokens),
  • Mistral Small 4: $0.0006 (710 tokens),
  • Ministral 3 8B: $0.0003 (883 tokens).

At these amounts, API cost becomes irrelevant for personal use. The $10 credit from the free Vibe plan is enough for several thousand similar queries to Medium 3.5. It only matters at scale in automation, for example when summarizing thousands of documents daily.

When You Don’t Need to Pay at All

  • A few questions per day: Free plan in browser or app.
  • Privacy and your own hardware: models under Apache 2.0 license (Ministral 3, Mistral Small 3.2 and 4, Large 3) can be downloaded and run for free, including commercially. Ministral 3 8B takes 6 GB in Ollama and runs on a card with 8 GB memory. We show how to do this in our Ollama guide.
  • API testing: free $10 monthly credits from the Free plan.
  • Students: Student plan at $5.99 gives more API credits ($30) than Pro.

Frequently Asked Questions

How much does Mistral AI cost?

Mistral Vibe chat is free in the basic version. Vibe Pro costs $14.99 per month on mistral.ai. API costs from $0.10 per million tokens (Ministral 3 3B) to $7.50 per million output tokens (Medium 3.5).

Is Le Chat free?

Yes. Le Chat, now Mistral Vibe, has a Free plan with limited messages, web search, connectors, and $10 API credits per month.

Is Mistral API cheaper than OpenAI?

Lower and mid-tier models yes: Mistral Small 4 costs $0.15 and $0.60, and Ministral 3 8B costs $0.15 per million tokens. With OpenAI’s cheaper models the differences are small, though, and you need to compare specific model pairs rather than companies.

Can I use Mistral models commercially for free?

Yes, if you run them yourself and they have an Apache 2.0 license (Ministral 3, Small 4, Large 3, Voxtral Small). Mistral Medium 3.5 has a modified MIT license with restrictions at large scale, and Codestral requires a paid license.

Is there a student discount?

Yes, the Student plan costs $5.99 per month instead of $14.99, after status verification and for a maximum of 12 months.

Google · Your sourcesQuantslant higher in your Google - one clickAdd to preferred sources →

Sources and Verification Date

Verified October 3, 2026.

Kinetyra

// read next

More on this from QUANTSLANT

Piotr Olszewski

Editor and publisher · Quantslant

I edit and publish Quantslant. AI tools help with research, drafts and translation; choosing topics, running the tests, checking facts and the decision to publish are mine, and so is the responsibility for what appears here.

read nextCursor Free Plan: What Hobby Offers and How to Use ItguidesCursor Free Plan: What Hobby Offers and How to Use ItCursor's free Hobby plan gives you the full editor with limited AI requests and Composer access. No credit card needed, but the Pro trial is gone.

Most read

  1. 1Best AI Video Generator Apps for Your Phone (2026)
  2. 2Claude Opus 5.5 Motion Graphics: 60s Showreel from One Prompt
  3. 3Wan 2.2 on RTX 4090: timings, VRAM and four test clips
// site map

This page in other languages6

Quantslant in numbers 1,427posts in archive26news from the last 7 days18video models in the observatory
🎬 AI Video Benchmarks: Higgsfield DoP vs OpenAI Sora • Higgsfield vs Runway Gen-3 • AI Video Credit Calculator
× ‹ enlarge ›