Skip to content
runcrate
Console

MISTRAL API

Every Mistral model, on one API key.

Call Mistral Large 3, Small 4, Devstral 2 and the Ministral family through the OpenAI chat completions format, billed per token at the prices below.

Get an API keySee model prices

Hand this to your agent

Add a Runcrate model to this project: run curl -fsSL"https://runcrate.ai/agents.md?intent=build"and follow it.

Your agent asks for an API key. Create a key

Works inClaude CodeCodexOpenCodeClineRoo CodeKilo CodeCursor
Models in the line10
Longest context256K tokens
Rate limitNone on paid models

Quick start

Structured output in one request. Or read the request it sends.

Your agent writes and runs it. The same key works for every model and every GPU.

Docs
Use Mistral Small 4 on Runcrate to turn tickets.csv into a JSON list of bugs and save it as bugs.json. Runs curl https://api.runcrate.ai/v1/chat/completions \ -H "Authorization: Bearer rc_live_YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "mistralai/devstral-2-123b", "messages": [{"role": "user", "content": "Write a merge sort in Go."}] }'.

Available models

Mistral's line, as the catalog lists it today. Billed prices, as charged.

ModelMakerPriceDetail
mistralai/mistral-large-3Mistral AI$0.525 in, $1.58 out per 1M tokensThe largest, for hard tasks and long documents. 256K context, tool calling, JSON schema output, image input
mistralai/mistral-small-4Mistral AI$0.158 in, $0.63 out per 1M tokensGeneral chat and agents. 256K context, tool calling, JSON schema output, image input
mistralai/devstral-2-123bMistral AI$0.42 in, $2.10 out per 1M tokensCode and repository agents. 256K context, tool calling, JSON schema output
mistralai/magistral-small-2509Mistral AI$0.525 in, $1.58 out per 1M tokensStep-by-step reasoning. 128K context, tool calling, JSON schema output, image input
mistralai/ministral-3-14b-instructMistral AI$0.21 in, $0.21 out per 1M tokensSmall and quick, for extraction and routing. 128K context, tool calling, JSON schema output, image input
mistralai/ministral-3-8b-instructMistral AI$0.158 in, $0.158 out per 1M tokensSmall and quick, for extraction and routing. 128K context, tool calling, JSON schema output, image input
mistralai/ministral-3-3b-instructMistral AI$0.105 in, $0.105 out per 1M tokensThe smallest, for classification at volume. 128K context, tool calling, JSON schema output, image input
mistralai/Voxtral-Small-24B-2507Mistral AI$0.00315 per minuteSpeech to text on /v1/audio/transcriptions
mistralai/Mistral-Nemo-Instruct-2407Mistral AI$0.0305 in, $0.0315 out per 1M tokens128K context, tool calling, JSON schema output
mistralai/Mistral-Small-3.2-24B-Instruct-2506Mistral AI$0.0788 in, $0.21 out per 1M tokens128K context, tool calling, JSON schema output, image input
See every model price

Why Runcrate

Built for production. One key, one balance.

  • Pick the size per job

    Ministral 3B, 8B and 14B for classification, extraction and routing at volume; Mistral Small 4 and Large 3 for agents and long documents. Each model's context window is in the table.

  • Devstral 2 for code

    Devstral 2 123B is Mistral's coding model, with tool calling, for agents that read and change a repository.

  • Tools and JSON schema

    The Mistral chat models here take the OpenAI `tools` parameter and `response_format` with a JSON schema, so a reply parses without retries.

  • Voxtral on the same key

    Transcribe audio with Voxtral Small on /v1/audio/transcriptions at $0.00315 per minute, billed from the same balance as the chat models.

FAQ

Common questions.

Related

Read next.