Mistral AI Pricing in 2026: Pro Costs, Free Tier Limits, and API Rates
Mistral Pro is $14.99/mo with a 150-message soft cap. Here's what the free tier actually gives you and what the API costs per token.
Mistral AI's pricing page tells you just enough to get confused. Three consumer tiers, a separate API billing system, and usage caps buried in fair-use language that most people never read. I spent the time pulling apart every tier so you don't have to.
Here's the full breakdown of what Mistral costs in 2026, from the free tier through Enterprise, plus the API pricing that runs on a completely separate billing track.
Stop Paying for Kong. You Have NGINX
NGINX handles a third of the internet. Most engineers use 10% of it. This book closes that gap.
Read the Config - $3.99Le Chat Free Tier: What You Actually Get
Mistral's free tier gives you access to Mistral Medium and Small models through Le Chat, their consumer chat interface. The daily cap sits around 25 messages per day, which is enough to kick the tires but will run dry fast if you're using it for real work.
What's included at zero cost:
- Access to Mistral Medium and Small models (not Large)
- Code interpreter for running Python in-chat
- Document uploads for analysis
- AFP-verified news search integration
- Roughly 25 messages per day before you hit the wall
The free tier runs on the same inference engine as paid plans. Mistral doesn't throttle speed for free users: they just cap how many messages you can send. That's a meaningful difference from providers that degrade response quality on free tiers.
What you don't get: access to Mistral Large, Flash Answers, No Telemetry Mode, or the expanded daily limits. If any of those matter to your workflow, you're looking at Pro.
Le Chat Pro: $14.99/Month
This is where most individual users land. At $14.99 per month, Mistral Pro is cheaper than ChatGPT Plus ($20/month) and Claude Pro ($20/month), making it the most affordable premium AI chat subscription from a major provider.
Here's what Pro unlocks:
- Access to all Mistral models, including Mistral Large
- 150 Flash Answers per day (the fast, concise response mode)
- Approximately 150 messages per day (a soft cap at roughly 6x the free tier)
- No Telemetry Mode: your prompts are never used for model training
- Up to 15GB of document storage
- Up to 1,000 projects for organizing conversations
- Mistral Vibe for all-day coding sessions (with pay-as-you-go beyond included usage)
- Chat and email support
The No Telemetry Mode is the headline feature for anyone working with sensitive material. Lawyers, journalists, developers working on proprietary code: that toggle means your data stays yours. Other providers offer similar privacy settings, but Mistral makes it a front-and-center selling point.
The "unlimited" messaging claim comes with a soft cap. You'll get about 150 messages per day under fair use. For most people, that's plenty. Heavy users running extended coding sessions or document analysis workflows might bump into it.
Student Plan: $7.04/Month
Mistral offers a student discount that cuts the Pro price by 53%. You need a .edu email address (or regional equivalent) plus ID verification, and the discount requires annual renewal.
You get the same feature set as Pro. The only difference is the price. If you qualify, this is the cheapest path to a premium AI subscription from any major provider.
Team and Enterprise: Custom Pricing
Team pricing isn't publicly listed. You need to contact Mistral's sales team for a quote. Enterprise contracts typically start around $20,000 per month for organizations that need:
- Domain name verification for brand consistency
- Shared RAG (retrieval-augmented generation) libraries at up to 30GB per user
- Admin console with consolidated billing
- Default data-training opt-out with zero configuration
- Access to Mistral Medium 3 at significantly lower token rates than competitors
- Private or on-premises deployment options
- No-code Agent Builder tools
For organizations already spending over $3,000 per month on OpenAI credits, Mistral's Enterprise tier often represents a significant cost reduction, especially for workloads that can route through Medium 3 instead of premium-tier models.
API Pricing: A Completely Separate System
This is where most people get tripped up. Le Chat Pro does not include API credits. The consumer subscription and the developer API are billed independently. Paying $14.99/month for Pro gives you zero API access.
Mistral's API runs through their La Plateforme service, and pricing is per-token:
| Model | Input (per 1M tokens) | Output (per 1M tokens) | Context Window |
|---|---|---|---|
| Mistral Large 3 | $0.50 | $2.00 | 131K tokens |
| Mistral Medium 3 | $0.40 | $2.00 | 131K tokens |
| Mistral Small 4 | $0.15 | $0.60 | 128K tokens |
| Devstral 2 | $0.40 | $2.00 | 128K tokens |
| Mistral Nemo | $0.02 | $0.06 | 128K tokens |
| Codestral | Varies | Varies | 32K tokens |
The API does have a free tier (sometimes called the "Experiment" tier) with rate limits. It's designed for testing and prototyping, not production workloads. If you're building something real, you'll need to add a payment method and pay per token.
Mistral Nemo at $0.02 per million input tokens is the cheapest model in their lineup and one of the most affordable API options from any major provider.
Le Chat Pro vs. Codestral: The Developer Gotcha
Codestral is available inside Le Chat if you have a Pro subscription: you can use it for coding assistance in the chat interface. But if you want to call Codestral through the API for IDE integration or automated workflows, that's billed separately at per-token API rates. The Pro subscription doesn't cover it.
This matters for developers who assumed a $14.99 subscription would cover their coding workflow end-to-end. It covers chat-based coding help. It does not cover API-driven code generation.
How Mistral Compares on Price
A quick comparison against the other major consumer AI subscriptions:
| Provider | Monthly Price | Key Differentiator |
|---|---|---|
| Mistral Le Chat Pro | $14.99 | Cheapest premium tier, No Telemetry Mode |
| ChatGPT Plus | $20.00 | Broadest feature set, DALL-E integration |
| Claude Pro | $20.00 | Best for long documents and reasoning |
| Google Gemini Advanced | $19.99 | Deep Google Workspace integration |
| Perplexity Pro | $20.00 | Best for cited research |
Mistral undercuts every major competitor by $5 or more per month. The tradeoff is a smaller ecosystem: no image generation, no native mobile voice mode (as of early 2026), and a less mature plugin system. But for pure text-based AI work with strong privacy controls, it's hard to argue with the price.
Who Should Pay for Mistral Pro
Mistral Pro makes the most sense for three groups:
Privacy-conscious professionals. The No Telemetry Mode is a genuine differentiator, not marketing. If you work with client data, legal documents, or proprietary code and want contractual assurance that your prompts aren't training the model, Mistral offers the clearest guarantee at the lowest price.
European users and organizations. Mistral operates entirely within European jurisdiction with servers in Paris. For GDPR compliance and data residency requirements, this is the simplest option. No need to evaluate data processing agreements or worry about transatlantic data transfers.
Budget-conscious power users. If you're paying for ChatGPT Plus primarily for text-based work and don't use DALL-E, voice mode, or the plugin ecosystem, switching to Mistral saves $60 per year with comparable model quality for most tasks.
Who Should Skip It
If you need image generation, integrated voice conversations, or deep integration with an existing ecosystem (Microsoft 365, Google Workspace), Mistral isn't there yet. The product is laser-focused on text: chat, code, and document analysis. That focus keeps the price low, but it means you might still need a second subscription for specialized tasks.
The Bottom Line
Mistral's pricing in 2026 is straightforward once you understand the split between consumer and API billing. Le Chat Pro at $14.99 per month is the cheapest premium AI chat subscription from a major provider. The API is competitively priced and runs independently. The free tier is genuine but limited to about 25 messages per day on mid-tier models.
The real question isn't whether Mistral is affordable: it is. The question is whether the privacy-first, text-focused feature set matches your workflow. If it does, you're getting a premium AI assistant for less than most people pay for a streaming service.