YearlyRank AI Tool2026AI tools listRanking
⌘K

Getting started with Mistral models in Studio or the API

· Tutorial

This tutorial explains how to start with Mistral’s official models using pages retrieved on 16 November 2026: the Mistral home page, Models Overview, API pricing, the quickstarts for creating an API key and sending a first request, Billing, and Create your organization. Confirm current steps and prices on Mistral’s docs before you add a payment method or generate a key. This guide does not assign a YearlyRank official score. See Mistral and methodology. YearlyRank lists Mistral as a model family, not as the Mistral Vibe chat app (that catalogue entry is separate; see How Mistral Vibe, the Gemini app, and Microsoft Copilot differ).

Purpose. Official documentation, “Models Overview,” says Mistral develops, or makes available, open-weight and commercial large language models. Official featured models on that page include Mistral Large 4, described there as an open-weight, general-purpose multimodal model, and Mistral Small 4. The first job is to create a Mistral account, activate billing so API keys work, create a key in Studio, and send a first official API request. Official docs also point to the Hugging Face organization huggingface.co/mistralai as a place to obtain models for self-hosted use; this guide does not invent install commands for that path.

Steps

1. Create a Mistral account and, if you are setting up a team, an organization. Official “Create your organization” says to start with a Mistral AI account (the docs link account creation to console.mistral.ai). Log in to Studio. Official help says new accounts are prompted to name their organization and are redirected to Admin automatically; if you already have an account without an organization, open Admin and follow the setup prompt. A default workspace is created automatically; official help says you can skip creating another if one is enough.

2. Activate billing before you rely on an API key. Official organization and API-key quickstarts say a plan with API access and a payment method are required, and that team members cannot generate working API keys until billing is active. Official steps: open Admin Panel → Subscriptions → Billing, click Add payment method, and connect your card or a cloud marketplace account (official prerequisites name a credit card or a cloud billing account on AWS or Azure Marketplace). You can optionally set a Monthly spending limit before saving. Official help says billing takes 2–3 minutes to take effect; if a key appears inactive right after, wait and retry.

Official Billing docs also describe prepaid credits, which are applied to subscription invoices and are required to enable pay-as-you-go. You can add credits or redeem a gift code, and you can turn on Auto Recharge so that when the balance falls below a threshold you set, Mistral charges your payment method. Official help says that to use the API and Vibe Code beyond included monthly usage, the organization needs both pay-as-you-go enabled and a positive credit balance. When the balance reaches zero, official help says API and Vibe Code requests beyond included monthly usage fail with HTTP status 402, and you add credits from the link in the error message. Credits are deducted including tax, while spending limits apply before tax, so credits can run out before you reach a spending limit.

3. Create an API key in Studio and store it. Official quickstart, “create your first API key,” takes about five minutes. Official steps:

  1. Open the Studio console and go to API Keys.
  2. Select New key.
  3. Enter a Name that identifies the key (the docs use quickstart as an example).
  4. Set an Expiration date.
  5. Choose a Connector access scope: Shared connectors only, or Private and shared connectors.
  6. Create the key, then copy it and store it in a password manager or secrets vault. Official help warns that the full key appears only once and cannot be retrieved after you close the confirmation dialog.

Official docs then set the key in the current terminal session:

export MISTRAL_API_KEY="your-api-key"

The Mistral SDKs and the examples in the docs read that variable. Official help says to set an expiration date on every key, rotate keys regularly, and delete a key as soon as you stop using it or suspect it has leaked.

4. Send a first official request. Official key quickstart includes this test curl to https://api.mistral.ai/v1/chat/completions, using the model id mistral-small-latest and a user message “Hello, Mistral!”:

curl https://api.mistral.ai/v1/chat/completions \ -H "Authorization: Bearer $MISTRAL_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "mistral-small-latest", "messages": [{"role": "user", "content": "Hello, Mistral!"}] }'

Official help says you should see a JSON response with the model’s reply in choices[0].message.content, plus id, created, and usage. If you get {"detail":"Invalid API Key"} with HTTP 401, official help says to check that MISTRAL_API_KEY holds the full key in the same terminal. Official “send your first API request” repeats the same model id and also documents a Python path: Python 3.10 or later and uv; uv init first-request --python 3.12 and uv add mistralai; then a short script using Mistral from mistralai.client and client.chat.complete. A 429 error, official help says, means you hit a rate limit — wait, then retry with exponential backoff. Official docs also link a “Compare models in the playground” quickstart if you want a Studio UI path; follow that article rather than clicks invented here.

Plan and pricing notes

Official API pricing retrieved on 16 November 2026 lists prices by family, in amounts per million tokens unless a row says otherwise. Clearly labelled generalist figures that day included:

  • Mistral Small 4: $0.15 input, $0.015 cached input, $0.6 output per million tokens.
  • Mistral Large 4: $0.7 input (also shown as $0.7 /M Tokens on the model heading), $0.07 cached input, $2.1 output per million tokens.
  • Ministral 3 14B: $0.2 / $0.02 / $0.2; Ministral 3 8B: $0.15 / $0.015 / $0.15; Ministral 3 3B: $0.1 / $0.01 / $0.1.
  • Codestral: $0.3 / $0.03 / $0.9.

OCR rows on the same page used a different unit: OCR 4.1 was listed at $4 and $0.4, and official text labelled those OCR prices per 1,000 pages. Official help lists Mistral Moderation 2 as Free, and self-hosted open-weight models such as Shieldstral 1.0 and Timestral 1.0 with the note that costs depend on your own infrastructure. For any other model, including third-party models listed on the same page, use the live pricing table.

There is no consumer “Free / Plus / Pro” chat-plan table on these model docs. Billing is organization-level, with a payment method, optional monthly spending limits, and prepaid credits as official Billing help describes. Admin or Billing organization roles manage billing.

Who it fits

This start path fits someone who wants official Mistral models through Mistral Studio and the API, and who can add a payment method so keys will work. It is a model-and-API guide, not a consumer chat-app guide for Mistral Vibe. Open-weight downloads via Hugging Face are pointed to, not walked through, because this tutorial does not invent host-side install steps.

Closing takeaway

Create an account at console.mistral.ai, activate billing in Admin, generate a key under API Keys, export MISTRAL_API_KEY, and send the official mistral-small-latest test request. Official prices above are the 16 November 2026 docs figures per million tokens (OCR excepted). For the current table, rate limits, and credit rules, use the live API pricing and Billing pages.

Sources

Retrieved on 16 November 2026 (Asia/Shanghai):

  1. Mistral AI, home page, https://mistral.ai
  2. Mistral Docs, “Models Overview,” https://docs.mistral.ai/models
  3. Mistral Docs, “API pricing,” https://docs.mistral.ai/inference/pricing
  4. Mistral Docs, “Quickstart: create your first API key,” https://docs.mistral.ai/getting-started/quickstarts/studio/activate-and-generate-api-key
  5. Mistral Docs, “Quickstart: send your first API request,” https://docs.mistral.ai/inference/quickstarts/send-your-first-request
  6. Mistral Docs, “Create your organization,” https://docs.mistral.ai/getting-started/quickstarts/admin/create-organization
  7. Mistral Docs, “Billing,” https://docs.mistral.ai/admin/billing-usage/billing
  8. Mistral AI, Studio / console, https://console.mistral.ai
  9. Hugging Face, mistralai organization, https://huggingface.co/mistralai