Skip to main content

How do API rate limits work and how can I increase them?

Updated this week

API rate limits

To keep the service reliable and fair, we cap how much each Organization (including all its Workspaces) can use the Mistral APIs.

We enforce three types of limits:

  • Requests per second (RPS)

  • Tokens per minute

  • Tokens per month

Rate limits are set at the Organization level. Token limits include both input and output tokens.

πŸ“Œ Your limits are determined by your usage tier. Each tier comes with its own thresholds. See the table below for details.

Usage tiers

You can see the rate and usage limits for your Organization in the Limits section of la Plateforme.

La Plateforme offers several tiers, including a free API tier with conservative rate limits.

πŸ”‘ The free API tier is intended for evaluation and prototyping purposes. For actual projects and production workloads, we recommend upgrading to a higher tier.

How can I increase my rate limits?

Your limits may increase when you cross these total-billing thresholds:

Total billed amount*

Previous tier

New tier

> $20 / 20 €

Tier 1

Tier 2 (automatic upgrade)

> $100 / 100 €

Tier 2

Tier 3 (automatic upgrade)

> $500 / 500 €

Tier 3

Tier 4 (automatic upgrade)

> $2000 / 2000 €

Tier 4

* Total billed amount is the cumulative sum of all invoices.

πŸ“© Needs higher limits? Please contact our support team, providing the following details about your intended use case:

  • specify the RPS you are aiming for

  • provide details about your intended usage including the specific model you plan to use

  • give us an approximate estimate of the number of tokens required per minute and per month

Did this answer your question?