Skip to content

Mistral AI

Mistral is a Paris-based AI company that offers open-weight and commercial language models, a model customization service called Forge, an agent-building platform called Studio, and the Vibe assistant for research and coding. It targets organizations that want to run frontier AI on their own terms,

Free / Pay-per-useOpen Models & Fine-Tuning
70

Overview

Mistral is a French artificial intelligence company that builds open-weight and commercial language models, along with the tooling to customize, deploy, and use them. Its pitch rests on control: organizations can adopt frontier models without handing every dependency to a single closed vendor.

Open-Weight Models as a Starting Point

Mistral's catalog centers on frontier and specialist open-weight models. Teams can run them unchanged or adapt them to a particular domain. The appeal for many buyers is that open weights allow self-hosting, which matters when data residency, procurement rules, or long-term vendor risk are part of the decision. Mistral notes that its open-weight models, such as Mistral 7B, carry an Apache 2.0 license for research and individual use, while commercial deployments require a Mistral license with separate terms covering derivatives and production use.

Forge and Studio for Custom Work

Two products extend the base models. Forge lets organizations customize models around their own data, use cases, and requirements, with the resulting intelligence treated as the customer's own. Studio handles the operational side: building, governing, and improving agents and applications. Together they cover a path from a general-purpose model to something tuned for a specific business process, with governance built into the application layer rather than bolted on afterward.

Vibe as the Everyday Interface

Vibe is the end-user layer, bringing the models into research, coding, and routine tasks. It runs on the web, on iOS and Android, in the command line, and in the IDE. Paid tiers add all-day coding sessions, remote coding agents, hooks, and agent observability. For developers who want an intelligent AI coding assistant tied to the same model stack they deploy, Vibe is the entry point, and the Free plan offers a limited version of it alongside access to Studio.

Deployment and Compute Choices

Mistral supports on-premises installation, deployment inside a customer VPC, or fully managed hosting. Inference can stay in-region, and customers can reserve dedicated GPU capacity for predictable throughput without managing hardware themselves. That range is unusual among model vendors, many of which offer only an API. It also means the deployment decision can follow existing security and infrastructure practices rather than forcing a migration.

Plans and API Pricing

The Free plan provides limited access to Vibe and Studio. Pro costs $14.99 per month, or $5.99 for verified students, and adds higher limits, faster responses, all-day coding, spreadsheet and slide generation, and API credits. Team runs $24.99 per user per month with higher ceilings, up to 30GB of storage per user, domain verification, and data export. Enterprise starts at $50 per month and adds custom models, agents, and workflows, audit logs, SAML SSO, and white labeling. API pricing is mostly per million tokens, with input and output counted separately; Mistral Large is listed at $0.5 per million input tokens and $1.5 per million output tokens. Batch processing halves the price, and cached input tokens cut input cost by up to 90%. OCR, speech, and tool APIs use their own units.

Where the Fit Is Strongest

Mistral makes the most sense for organizations that treat model choice as an infrastructure decision. Regulated industries, European enterprises with data-sovereignty obligations, and platform teams that want to fine-tune on proprietary data all have clear reasons to look here. Smaller teams may find the tier structure and licensing terms worth studying before committing, since commercial use of open-weight models carries its own agreement. The company's customer roster, which includes large insurers, banks, and industrial firms, suggests the enterprise motion is well established.

Core features

  • Open-weight frontier and specialist language models that can be used as-is or adapted
  • Forge for customizing models around an organization's own data and requirements
  • Studio for building, governing, and improving agents and applications
  • Vibe assistant across web, mobile, CLI, and IDE for research, coding, and everyday tasks
  • Deployment options spanning on-premises, a customer VPC, or Mistral-managed hosting
  • Regional inference and reserved dedicated GPU capacity for predictable throughput
  • Connectors, workflows, task scheduling, and document libraries in the Vibe plans
  • Team and enterprise controls including SAML SSO, audit logs, data export, and white labeling

Use cases

  1. 1Running open-weight models in a private environment where data residency matters
  2. 2Fine-tuning or customizing models on proprietary data through Forge
  3. 3Building and governing internal agents and applications with Studio
  4. 4Agentic coding in the terminal, IDE, or remotely through Vibe
  5. 5Reserving dedicated GPU capacity for steady production inference
  6. 6Standardizing a collaborative AI workspace across a team with SSO and audit logs

Strengths and things to check

Strengths

  • Open-weight models give teams more control over hosting and customization than closed APIs alone
  • Deployment flexibility covers on-premises, VPC, and managed options with in-region inference
  • The stack spans base models, customization, agent tooling, and an end-user assistant
  • Enterprise features such as SSO, audit logs, and white labeling are available on higher tiers
  • Published per-token pricing and batch discounts make API costs easier to model

Things to check

  • Commercial deployment of open-weight models requires a Mistral license with separate terms for derivatives and production use
  • Several capabilities, including custom deployments and white labeling, sit behind Enterprise pricing that requires contacting sales
  • Model training is on by default on paid plans unless users opt out
  • Feature availability varies widely by tier, so plan comparisons need a close read

Pricing

Mistral publishes a Free plan with limited Vibe and Studio access, Pro at $14.99 per month (students $5.99), and Team at $24.99 per user per month. Pro includes $25.5 per month in API credits on one listing and $15 per month on another. Enterprise starts at $50 per month for private deployments with custom models, agents, and workflows. API models are mostly billed per million tokens, with Mistral Large at $0.5 per million input tokens and $1.5 per million output tokens; batch processing cuts prices by 50%, and cached input tokens reduce input cost by up to 90%. OCR is priced per 1,000 pages, speech models per minute, and tool APIs per call. Prices are subject to change.

Basic information

Website
mistral.ai
Added
Oct 8, 2026
Platforms
Web
Chinese & English
Not marked
API
Not marked
Self-hosting
Not marked
Best for
Mistral suits engineering, platform, and AI operations teams at organizations that need open-weight models, in-region inference, or private deployment, along with developers who want a coding and research assistant backed by the same model stack.
Information updated
Oct 16, 2026
Website status
Checking
Corrections · Privacy · Data requests?
Contact us

Features, pricing and availability may change. Always verify details on the official website. Maintained by steven · Last updated Oct 16, 2026 →

Related tools

All categories