Skip to content
MyStackPlan
LLM APIs & Inference

Mistral AI API pricing

Mistral's models by API, billed per token. Includes small, cheap models and open-weight options.

Try it in the builderOfficial pricing pageReport a wrong price

Every plan below was checked against Mistral AI API's own pages on . Prices in USD.

Plans

Pay as you go

Pay as you go

At the limit
Bills for extra usage
  • Billed per million tokens, with input and output priced separately.
  • Batch processing is 50% cheaper, and cached input tokens cost up to 90% less on repeated prompts (per Mistral's pricing FAQ).
  • OCR is priced per 1,000 pages, speech per minute and tool APIs per call. Those are not in this estimate.
  • A Free mode lets you create API keys and use included monthly usage within limits shown on the Limits page of your Mistral Admin Panel. Mistral does not publish those numbers (https://docs.mistral.ai/admin/billing-usage/usage-limits), so Free mode is not listed as a plan here.
  • In Free mode Mistral may use your API inputs and outputs to train its models by default. Free and pay-as-you-go users can both opt out at any time (https://help.mistral.ai/en/articles/347617-do-you-use-my-user-data-to-train-your-artificial-intelligence-models).

Source: docs.mistral.ai · checked

Models

ModelInput per 1M tokensOutput per 1M tokensAnnounced change
Ministral 3 3BSmallest and cheapest, open weights (Apache 2.0)$0.10$0.10None announced
Ministral 3 8BOpen weights (Apache 2.0)$0.15$0.15None announced
Mistral Small 4Instruct, reasoning and coding in one model$0.15$0.60None announced
Mistral Large 3Open-weight multimodal model$0.50$1.50None announced
Mistral Large 4Public preview. Shown at a sale price; the model page lists the original price as $1.36 in and $4.18 out and gives no end date for the sale$0.68$2.09None announced
Mistral Medium 3.5Frontier-class model for coding and agent tasks$1.50$7.50None announced

Alternatives for llm apis & inference